THESE Date 0-7639 This is to certify that the thesis entitled Choice of land preparation techniques for rice cultivation in Indonesia The utilization of small two wheeled tractors at the farm level in Karawang and Subang counties presented by in the wet season of 1980 Soedjatmiko has been accepted towards fulfillment of the requirements for the—EhD—___ degree in Jean—Eng . Technology Major professor 6/2 37/8! 4—---I -11‘1— . MSU RETURNING MATERIALS: Place in book drop to remove this checkout from LIBRARIES 1...¢,...._ your record. FINES will be charged if book is returned after the date _P,p1 stamped below. e '2: '3343 ‘ 1" 5 'J: ”67'? 5_ “r E T: If ? . CHOICE OF LAND PREPARATION TECHNIQUES FOR RICE CULTIVATION IN INDONESIA THE UTILIZATION OF SMALL TWO WHEELED TRACTORS AT THE FARM LEVEL IN KARAWANG AND SUBANG COUNTIES IN THE WET SEASON OF 1980 By Soedjatmiko A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Engineering 1981 ABSTRACT CHOICE OF LAND PREPARATION TECHNIQUES FOR RICE CULTIVATION IN INDONESIA THE UTILIZATION OF SMALL TWO WHEELED TRACTORS AT THE FARM LEVEL IN KARAWANG AND SUBANG COUNTIES IN THE WET SEASON OF 1980 By Soedjatmiko Insufficient food production and the ever increasing demand for rice has caused the Government of Indonesia to designate§>rice as a strategic commodity. In its attempt to stimulate an increase in rice production the Government of Indonesia has undertaken two main efforts: 1) the expansion of newly developed agricultural land, and 2) the inten- *\1 sification of production of the existing rice land. In association with this program, the Government is currently operating several selective agricultural mechanization projects (including tractorization) to help small farmers. The introduction of small two wheeled tractors in densely populated areas, such as West Java, has generated considerable debate, about private benefits versus the possible adverse social impacts. This research focused on the choice of rice land preparation techniques during the 1980 wet season in the Northern coastal plains of West Java. There were three specific objectives: 1. To evaluate the operational performance of small two wheeled tractors for rice land preparation through direct field measurement techniques. Soedjatmiko 2. To evaluate the small tractor labor utilization and cost factors for rice land preparation. 3. To develop and test a systems analysis model to evaluate and compare rice land preparation techniques. A systems analysis approach was used as the analytical and problem evaluation technique. Model verification was conducted with direct measurement data from a stratified randomized design. Field data were analyzed using linear additive models, analysis of variance and linear regression methods. The simulation model consisted of farm level and tractor owners' subsystems. The significantconclusionsderived from the statistical and computer simulation analysis were as follows. Small two wheeled tractors greatly increased labor capacity for rice land preparation in the Northern coastal plains of West Java. The depth of the first manual labor tillage (6.2 cm) was significantly shallower than plowing (10.7 cm) with bullocks and tractor rototilling (10.3 cm). The difference of tillage depth between bullocks and trac- tor was not significant. Number of effective hours worked per day were significantly greater for tractors than for either manual labor or bullocks (the average was 11.6, 7.5 and 5.8 hours per day respectively). The simulated seasonal capacity of manual labor was 0.77 hectare, 3.4 hectares for bullocks and 27.4 hectares for tractors. The total labor utilized per hectare for land preparation was 4.1 man days with tractors, 10.6 man days with bullocks and 39.2 for human labor. The total labor input for rice preharvest activities was not significantly different Soedjatmiko for farms using tractors and those using bullocks. The average portion of rice land prepared in the villages studied was 15 percent by bullocks, 29 percent by manual labor and 46 percent by tractors. The farmers' cost for rice land preparation by bullocks was generally lower than by either manual labor or tractors. Manual labor cost was the highest at Rp 33,973/ha for two passes, and Rp 30,100 and Rp 26,600 per hectare for tractors and bullocks respectively. The tractor owners' cost for two passes of rototilling was Rp 12,800/ha and the simulated returns and costs ratio was estimated at 2.12. The performance of locally made tractors (IRRI type design) was not significantly different than for imported tractors. The returns and costs ratio of locally made tractors (2.46) was higher than for imported tractors (2.12), and the purchase price of Rp 1,650,000 in 1980 was 30 percent lower than the average imported tractors. Approved 774%; W Major Professor Approved Department Chairman ACKNOWLEDGEMENTS The author gratefully acknowledges the support of many individuals and institutions who contributed to this study. In particular the author wishes to express his appreciation to: Dr. Merle L. Esmay, the author's Major Professor, for his guidance, unfailing courtesy and patient encouragement during this work. Dr. Robert D. Stevens, Dr. C. Allan Rotz, and Dr. Robert H. Wilkinson who also served on the author's guidance committee, for their helpful suggestions and ideas in developing this dissertation and continued interest. Dr. Bill A. Stout for his strong recommendation to the Government of Indonesia which made it possible for the author to further study at Michigan State University. United States Agency for International Development, United States Department of Agriculture, The Ford Foundation, The Ministry of Agriculture of The Republic of Indonesia for their financial support. Ir Jafri Jamalludin, the Director of Food Crop Production, Ir Subagyo Wirjosoemarto, Head of The Subdirectorate of Agricultural Mechanization for their encouragement and support. Mr. Muljoto, Ir Handaka, Mr. Soetiardjo, Mr. Soetopo and Mr. Soeparmin for their unfailing help in field data collection. ii TABLE OF CONTENTS CHAPTER 1 - INTRODUCTION. . . . . . . . . . . . . . . . . . . . 1.1 The Problem of Mechanization. . . . . . . . 1.2 Objective of the Study. . . . . . . . . . . . . 1.3 Research Design and Methodology . . . . . . . . CHAPTER 2 - DEMOGRAPHIC CHARACTERISTICS OF KARAWANG AND SUBANG COUNTIES AND EXISTING LAND PREPARATION TECHNIQUES . 2.1 Characteristics of the Counties Studied . . . . . . . 2. 2 Rice Growing Technique. . . . . . . . 2. 3 Land Preparation Technique and Constraints. CHAPTER 3 - AGRICULTURAL DEVELOPMENT THEORY RELATED TO APPROPRIATE TECHNOLOGY AND PREVIOUS RESEARCH IN INDONESIA O O O C I O O C C O C C C O O O O O O O 3.1 Agricultural Development Theory Related to Technological Changes . . . . . . . . . . . . . 3.2 Review of Rice Production which Includes Agricultural Mechanization in Indonesia. . . . . . . . . . 3.3 Tractorization Experiences in Other Developing Countries in South Asia . . . . . . . . . . 3.4 Previous Agricultural Mechanization Research Works in Indonesia. . . . . . . . . . . . . . . . CHAPTER 4 - DATA AND ANALYSIS FRAMEWORK OF FARM SURVEY AND DIRECT MEASUREMENT. . . . . . . . . . . . . 4.1 Research Design . . . . 4.1.1 Villages Studied. . . . . . . . . . . 4.1.2 Sample Unit and Size. . . . . . . . . . 4.1.3 Statistical Analysis Model. . . . . . . 4.2 Relevant Criteria for Land Preparation Measurement and Data Obtained . . . . . . . . . . . . . . . . 4.3 Labor Use and Cost for Land Preparation Derived from Farm Survey. . .~. . . . . . . . . . . CHAPTER 5 - MODEL DEVELOPMENT AND SIMULATION FOR LAND PREPARATION TECHNIQUES. . . . . . . 5.1 A System Approach for Choice of Land Preparation Alternatives. . . . iii 13 15 l9 19 22 24 27 29 29 30 32 33 39 55 63 64 Page 5.1.1 Identification of System Components. . . . 64 5.1.2 System Linkages. . . . . . . . . . . . . . 67 5.2 Rice Land Preparation Model. . . . . . . . . . . . . . 69 5.2.1 Time, Cost and Seasonal Capacity Analysis . . . . . . . . . . . . . . . . . 73 5.2.2 Farm Level Subsystem Model . . . . . . . . 75 5.2.3 Tractor Owners' Subsystem Model. . . . . . 77 5.3 System Simulation for Land Preparation . . . . . . . . 79 5.3.1 System Simulation Inputs . . . . . . . . . 82 5.3.2 System Simulation Outputs and Discussion . . . . . . . . . . . . . . . . 83 CHAPTER 6 - CONCLUSIONS. . . . . . . . . . . . . . . . . . . . . 98 APPENDIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 iv .10 .11 .12 .13 .14 .15 .16 .17 .18 LIST OF TABLES Demographic characteristics of Java relative to IndoneSia I Q I O O O O I O O O O O O O O O O O O O 0 Demographic features of Karawang and Subang counties . Demographic characteristics of villages under study. . Analysis of variance for choice of land preparation. Average age of land preparation operators. . . . Operator experience in land preparation. . . . . . . . Hours worked per day on the first tillage operation. . Hours worked per day on the second tillage operation Depth of tillage and soil mechanical properties. Plot size and percentage of untilled land. . . . . . Travelling speed of bullocks and tractors. Average of time required for the first tillage of one hectare of rice land . . . . . . . . . . . . . . . . . Average of time required for the second tillage of one hectare of rice land . . . . . . . . . . . . . . Average fuel and oil consumption for tillage operation Field performance data on locally made tractors. . . . Estimated price of equipment . . . . . . . . . . . . . Operator wages for manual labor, bullocks and tractors Land holding by sample farms Labor input for rice preharvest activities The amount of fertilizers and chemicals used on sample rice farms 0 O O O O O O O O O O 0 0 O O O O O O O O Page 10 31 37 4O 41 42 43 45 48 50 51 52 53 . 54 . 56 . 57 58 59 61 .19 .10 .ll .12 .13 .14 Level of seed planted and rice yield of the sample farms . Annual rate of change of components for the financial analysis . . . . . . . . . . . . . . . . . . . . . Examples of input values for sensitivity analyses Probability of delay, wait and heavy rainfall . . . . . . Simulated effective working hours in the wet season of 1980 O O O O O O O I I O O O O O O O O O O O O O O O O O 0 Simulated capacity of manual labor, bullock and tractor technologies for the wet season of 1980 (two passes of tillage OperatiOnS). o o o o o o o o o o o o o o o o o o 0 Simulated proportion of rice land prepared by manual labor, bullocks and tractors, wet season, 1980 . . . . . . Simulated amount of rice land prepared per man day (two passes of tillage) . . . . . . . . . . . . . . Simulated total labor utilization for two tillage operations 0 O I O O O I O O I O I O O O O O O O O I O O 0 Simulated hourly total costs of land preparation in the wet season of 1980 . . . . . . . . . . . . . . . . . . . . Simulated costs of land preparation per hectare (two passes, wet season, 1980). . . . . . . . . . . . . Tractor owners' earnings, costs and profitability index Simulated locally manufactured and imported tractor performance for rice land preparation (two passes) . Example of simulated waiting days for manual labor and animal power and least cost opportunity . . . . . . . The sensitivity analysis results . . . . . . . . . . . . . vi Page 62 80 81 82 84 85 86 88 89 90 91 92 93 95 96 LIST OF FIGURES Geographic situation of the studied areas. . . Preportion of rice land (2) transplanted each month in Cilamaya and Jatisari Districts . . . . . . Procedure for statistical analysis . . . . . . . Soil hardness profiles . . . . . . . . . . . . . . Blackbox diagram for choice of land preparation techniques at the farm and village levels. Simplified causal loops model for choice of land preparation techniques at the farm level . . . . Simplified flow chart for land preparation model at the farm level . . . . . . . . . vii Page 12 15 34 46 65 66 71 CHAPTER 1 INTRODUCTION Indonesia, a nation of 142 million inhabitants, consists of 3,000 islands stretching along the equator from 95 to 145. degrees east longitude. The total area is a little more than 1.9; million square kilometers or about five times the area of the State of Texas. With a population density of 679/km2, Java is the most densely popu- lated island in the world. About 90 million people, about 63 percent of the Indonesian population, live in an area that is less than 7 percent of the total land area (Central Bureau of Statistics, 1979). Agriculture, in this oil producing country, remains a dominant sector. In 1974, agriculture shared 39 percent of the gross national product; however, it gradually decreased to 34 percent in 1978 (Bureau of Statistics, Jakarta, 1978). The World Bank data indicate that in- come per capita per year was $220 in 1970, with a yearly growth rate of 3.5 percent from 1970 to 1975 (The World Bank Atlas, 1977). This growth of per capita income in the last decade has caused peOple to shift their food preference from lower nutrient foods to rice as their main staple (Ministry of Agriculture, 1977). The insufficient increase food production and the ever increasing demand for rice caused the Government of Indonesia to regard rice a strategic commodity. Rice production has thus become high priority in the national development program (Ministry of Agriculture, 1977). The focus of this thesis is on the contribution of agricultural technology, specifically tractorization, to the national development process, in particular rice production intensification. Since mechani- zation is generally categorized as being labor saving and capital intensive, the question arises whether tractorization is an appropriate technology to be introduced in such densely populated, low per capita income regions. This chapter presents l) a description of the problem regarding mechanization, 2) an identification of related research, and 3) the research objectives and methodology. 1.1 The Problem of Mechanization Inadequate food production, unemployment and uneven geographic distribution of the population are urgent problems in Indonesia. The Government has stressed the transmigration program and the intensifica- tion of agriculture to meet the ever increasing demand for food, in particular rice. The implementation of the intensive rice production program involved many of the small farmers (land holding less than 1 hectare), as they constitute almost 60 percent of the total population of Indonesia (Ministry of Agriculture, 1977; Central Bureau of Statis- tics, 1978). The Government is currently operating several selective agricultural mechanization projects to help these small farmers. In the coming decade these projects are to spread throughout the country. One project, the introduction of small (two wheel) tractors of 7 to 8.5 horsepower, for rice land preparation on small farms, has generated con- siderable debate over private benefits versus the possible adverse social impacts; while agricultural mechanization may increase land and labor productivity, reduce production cost and drudgery, it also has the potential of increasing the income disparities and unemploy- ment. This is particularly undesirable in the labor surplus areas of Indonesia. Furthermore, all of the small tractors introduced to date in Western Java have been imported. Their technical appropriateness for specific local conditions needs to be evaluated. Questions arise regarding the sophistication of equipment design, local repairability, durability, availability of spare parts and high initial cost. A critical question needing investigation is whether a small two wheel tractor of simple design (for example, the IRRI1 designed tractor) can be locally manufactured for a lower initial cost and with an acceptable technical and economic performance. In the long run, agricultural mechanization should not be dependent on importation. Past Government mechanization introduction projects have largely been based upon experimental data obtained from controlled conditions and less reliable secondary information. The present study was de- signed to provide data on small tractor performance, labor utilization, and the cost of rice land preparation on farmers' fields. Subsequently, these field performance data were used to verify a systems model de- veloped to compare and evaluate human, animal and tractor performance in rice land preparation. 1International Rice Research Institute. 1.2 Objectives of the Study The present research considered the choice of technology for rice land preparation during the wet season (Monson) in the Northern coastal plains of West Java. The study had two specific objectives: 1. Evaluate the operational performance of two designs of two wheeled small tractors for rice land preparation through direct field measurement techniques. 2. Compare small tractor, animal, and manual land preparation with focus on labor utilization and cost factors for rice land preparation. 3. Develop and test a system analysis model to evaluate rice land preparation techniques. Small tractor in this thesis is defined as 7-8.5 horsepower, two wheel, both imported (Japanese) and locally manufactured (IRRI design). This research is limited to the Northern coastal plains of West Java due to resource constraints. Further discussion regarding this speci- fic research location will be presented in Chapter 2. This research differs from most previous research on agricultural mechanization in Indonesia in two ways: 1) primary data are assessed by direct measurement at the farm level under farmer Operational con- ditions, and 2) following suggestions by the Michigan State University Task Force on Farming Systems Research, the small farmer values and goals are considered to explain and analyze preferences for land preparation techniques under specific ecosystems and locations (Norman 31331., 1980). 1.3 Research Design and Methodology A systems analysis approach is used as the analytical and problem evaluation technique. Model verification is conducted with data ob- tained by direct measurement in a stratified randomized design. Pur— posive sampling was conducted to select eight villages in Karawang and Subang counties as research sites. Random sampling was the method used to determine the farm survey and field measurement samples. The sample size was 216 farm households and 432 rice fields, located in eight villages in Karawang and Subang counties. The samples were stratified into three categories: 1) farms using small tractors, 2) farms using bullocks, and 3) farms using manual labor for land preparation. Field data were analyzed using linear additive models, analyses of variance and linear regression methods. The resulting generalized data were used for verification of the computer simulation model. Details of statistical analysis, systems model design and simulation modeling are presented in Chapters 4 and 5. CHAPTER 2 DEMOGRAPHIC CHARACTERISTICS OF KARAWANG AND SUBANG COUNTIES AND EXISTING LAND PREPARATION TECHNIQUES In Indonesia, there are two patterns of agriculture. The first is constituted by some 1,800 large estates, which manage about 2.22 million hectare (M? land and grow export type crops. The second consists of 14.4 million small farms, which share 14.17 million hectares of land devoted to food crop production for domestic consumption (Central Bureau of Statistics, 1973). This present research deals primarily with the small farmers associated with rice production. In its attempt to stimulate an increase in rice production, the Government of Indonesia is undertaking two main efforts: 1) the expan- sion of newly developed agricultural land, and 2) the intensification of production of the existing rice land. As a large portion of the existing rice land is located on the island of Java, it is necessary to review demographic characteristics of Java in relation to Indonesia as a whole (see Table 2.1). The ever increasing population pressure (an increase of 42 percent between 1961 and 1980) on Java reduces the availability of agricultural land for food crop production as well as pastures. Within ten years (1961-1971) land for food crop production in Java decreased by 2.5 per- cent. During the same time the draft animal population decreased by 7.9 percent and declined another 4.3 percent between 1971 and 1980 in Table 2.1. Indonesia.1 Demographic characteristics of Java relative to Items Units 1961 1971 1980 Population 1000 persons Indonesia 97,019 118,368 142,179 Java 62,993 76,030 89,657 Density Persons/km2 Indonesia 51 62 75 Java 477 576 679 Draft animal 1000 units Indonesia n.a. 9,359 8,7352 Java 5,611 5,170 4,9462 Land for food crops 1000 ha Indonesia 12,844 14,168 n.a Java 5,647 5,505 n.a Milled rice 1000 metric tons Production Indonesia 8,268 10,499 13,4582 Java 4,803 6,455 8,1342 Import“ 1,064 503 1,9643 1Source: Central Bureau of Statistics, Jakarta. 216 1978. 31a 1977. l'From: Gaiser (1980). Indonesia (see Table 2.1). A study conducted by the Agroeconomic Institute over a longer period (1930-1976), showed that the decrease of draft animal population in Java to have seen 2.6 percent per year (9). These macro level data on population growth, which has a direct bearing on the increase of manual labor supply, the decrease of draft animals and hectarages of agricultural land resources in Java and their interrelations are treated as exogenous inputs for the computer simulation model of this research. To complement these macro data con- cerning Java, additional data were collected about the selected villages and farms of this study in the Northern coastal plain of West Java (see Table 2.2). 2.1 Characteristics of the Counties Studied The Northern coastal plain of West Java has been known for hundreds of years as a rice producing area. At the beginning of the 17th century, Sultan Agung, the King of Mataram (Jogyakarta) commanded Bupati Wira Perbangsa to develop rice belts surrounding Batavia (the colonial name for Jakarta). The purpose was to provide sufficient "live" food storage for the Mataram army to defend the kingdom's terri— tory against the Dutch colonial expansion (Demography of Karawang county, 1979). Rice production systems have remained essentially un- changed since that time until the advent of the green revolution.1 However, this historical importance of the area is not the main reason for selecting the Northern coastal plains of West Java as the research site. The relevant considerations for the location of the study were: 1The green revolution is defined as a period in which modern science and technology begin to have major impact on agriculture and resulted in large continuing increases in land and labor productivity (Stevens, 1975). 1. West Java currently contributes greatly to the national rice supply. This province in 1978 produced 2.9 million tons of milled rice or about 22 percent of the total national rice production (Central Bureau of Statistics, 1978). 2. There was considerable increase in the use of small tractors during the last five years (Ministry of Agriculture, 1979). 3. The location was convenient for data acquisition in the relatively short period of time available for this project, without sacrificing essential require- ments for predetermined research design criteria as discussed in Chapter 4. Karawang and Subang counties are located in the middle of the Northern coastal plain of West Java. Eight villages from these two counties were selected as research sites. A demographic characteriza- tion of Karawang and Subang is presented in Table 2.2. Based on Table 2.2, it was calculated that each hectare of food crop land supports approximately nine individuals in Karawang or Subang counties without taking export of foodstuff of other counties into consideration. This figure might reach a level of 14 people per hec- tare by the year 2000. The increase of labor supply and the in-elastic demand for manual labor due to fixed or even declining land resources, unemployment may lower manual labor wages for land preparation. The Agroeconomic Survey Institute discovered that at the county level, approximately 65 percent of the total number of draft animals are in the productive age (between 3 to 12 years). Based upon this 10 Table 2.2. Demographic features of Karawang and Subang counties. Item Unit Karawang Subang Areas km2 1,504 2,051 Population Number Persons 1,109,044 942,500 Density Persons/km 738 460 Households Numbers 250,605 230,982 Agric. population Percent 55 87 Agric. labors2 Percent 15 27 Ag. male labors2 Percent 6 17 Pop./village (mean) Persons 9,902 5,891 Villages Numbers 112 160 Farm land Ha Total for food crops 133,210 107,209 Low land for rice 105,096 82,184 Intensification Percent 96 92 Draft animals2 Numbers 16,092 22,298 Small tractors2 Numbers 355 196 lKarawang and Subang county Bureau of Census and Statistics (1976). 2Agricultural Extension Service, West Java Province (1979). information and Table 2.2, it may be estimated that the ratio of potentially available draft animals for food production was 7.4 hec- tares/animal in Subang and 12.7 hectares/animal in Karawang (1979). The greater ratio in Karawang coincided with the greater number of tractors in this county as compared to Subang county. Important 'questions related to the choice of land preparation techniques are: l) the degree to which any increase of labor supply is evenly distri- buted between the urban and the rural areas; 2) the degree to which the economic structural transformation may shift manual laborers fast 11 enough from agricultural sector to non-agricultural sectors in the urban and industrial areas; and 3) the degree to which the green revolution stimulated alternate crOp management practices and brought about changes in the manual labor demand pattern for land preparation. The World Bank working paper (1979) reported that there were differential rates of pOpulation growth between urban and rural areas. For example in the decade of 1960 to 1970 the annual popula- tion growth in the urban areas was 4.2 percent compared to 2.3 percent in the rural areas. It was believed that this differential was due mainly to the out migration from the rural to the urban areas. 12 .mowwHH«> voavaum onu mo coaumsuwm ofisamuwoou .H.N ouswfim ”a momeHw> voquum may 2 mmoum HnwuumnbaH M _. a... $30 mam o ..2. ” hasnwfi=.lll\ ., _ mmoum unacamucso: 91me kuwmoo D I 103.1101... .6. ..... . 5’ 0.6.. C n b .151 1 .111 1 v _ . o m c.% m a " 0..“ " who.» £15m “ ...o 1 kg .112 13 2.2 Rice Growing Technique in Karawang and Subang Counties There are two seasons, wet and dry seasons, and three practices for growing rice in Karawang and Subang counties: 1) rice is grown under continuous flooded conditions, 2) "gogo" technique, where in all of rice growing stages are maintained under relatively dry conditions, and 3) "gogo ranca " technique, similar to gogo technique, but differs in its later (generative) stage, where rice field is turned into flooded condition until around two weeks before harvesting. The sub- merged (flooded) rice growing practice is the dominant method in the studied areas and accounts for more than 90 percent of the total rice land in Karawang and Subang counties (Agricultural Extension Service, West Java Province, 1980). This present study deals primarily with the flooded rice growing system. In Karawang and Subang counties, it is a common technique to per- form rice land preparation under submerged conditions, so that the rice field becomes soft enough for bullocks plowing or manual hoeing. The depth of water is usually around 5 to 8 cm. Rice seeds are soaked for two to three days and then sown in the seedbeds. Farmers are concerned about intensive care of rice nursery, for example, accurate water control, plant protection from birds, rat damage and other rice pests and disease attacks. Water level was raised to about 5 cm in the day time and drained in the night for aeration. The amount of land for the rice nursery was around 5 percent of the total rice field to be planted. Twenty-five to thirty kilograms of HYV rice seed was required for one hectare of rice field. The rice seedlings are kept in the nursery for twenty-one to twenty-five days for the best yield 14 (Farmers in the study areas, 1980). Land preparation must be done within this period (twenty-one to twenty-five days). Tilled rice land surface was maintained evenly and squared 25 by 25 cm2 or 20 x 30 cm2. Rice seedlings were pulled out by hand and transplanted (three seedlings) at the intersections of the squared lines. Basal fertilizers (nitrogen and phosphate) and top dressing (nitrogen) were applied just before transplanting and two to three weeks after transplanting respectively. Water level during rice grow- ing stages was maintained at about 5 to 10 cm deep and drained fre— quently one day before and after weeding and fertilizer applications. Finally, three weeks before harvesting, water was completely drained. Irrigation water supply was served for one liter per second for twenty- four hours continuously. One liter of active ingredient (pesticide) per hectare was sprayed once every fourteen days, depending on the intensity of rice pests and disease occurrence. Spraying was done using backpack compressed air Sprayers. Rice was harvested when three-fourths of the leaves and rice grains in the field looked yellowish, and cut by ani-ani (small hand knife) or by sickles. Threshing was carried out by foot treading or by beating rice stems on bamboo racks. After one or two days of dry— ing and cleaning, the rice was stored under village (17 percent to 18 percent) moisture content (Agricultural Extension Service, West Java, 1978). The typical amount of rice land transplanted monthly is illus- trated in Figure 2.2. The beginning of rice land preparation can be 15 100 _: 90 b 80 - 70 1 Dry season Wet season 60 1 50 _’ Jatisari Jatisari Cilamaya Cilamaya Pr0portional (Z) of Rice Land Planted 1 ll 1. [1 L / 1 .1 11 1_‘b 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec ---- Lbnths ---- Figure 2.2. Proportion of rice land (Z) transplanted each month in Cilamaya and Jatisari Districts. 1Total rice land in Cilamaya is 9,538 hectares and 9,488 hectares in Jatisari Distric. Figure 2.2 was processed based upon district level data (average of five years, 1976 to 1980). 16 estimated by shifting the transplanted curves three to four weeks ahead and the amount of land prepared monthly was proportional to the monthly transplanted areas. This information was based upon the average of five year records available at the district level in Jatisari and Cilamaya, where Jatisari, Jatiragas, Sukatani and Rawa Gempol villages are located. 2.3 Rice Land Preparation Techniques and Constraints in Karawang and Subang Counties Traditionally, rice land preparation was done by either manual labor hoeing or bullocks operated plowing and harrowing or sometimes by a combination of both manual labor and bullock tillage. Stout (1966) summarized the objectives of primary rice tillage as: 1) to loosen the soil in a depth of 10 to 15 cm; 2) to aerate the soil; 3) to initiate the cutting and distribution of dried organic matter, to kill grass and growing weeds and to create conditions favorable for decomposition of organic matter. These objectives were true in the studied areas. Rice land preparation in Karawang and Subang counties was accomplished by two times of manual hoeing or by bullocks plowing and two times harrowing. Two times of tractor roto- tilling was evaluated by farmers as producing the same result as bullock tillage and better than manual hoeing. Additional work related to rice land preparation was needed to make rice field surfaces more even for better water distribution, reshaping or reconstructing rice field levees and cleaning plant residue and weeds to be incorporated with the already tilled soil. 17 It was estimated by the farmers in the study areas that bullock capacity per season was around two and one half hectares and one half hectare for manual labor when intensive or good soil tillage was ex— pected. In the recent years, the last ten years in particular, farmers complained about the quality of rice tillage performed by either bullocks or manual labor, although they were paid on a daily basis, which should result in better tillage as compared to area based contract. For instance, shallower tillage, uneven plowing or some parts of the rice land was not tilled and sometimes land preparation was late which caused delay of transplanting. The delay of transplanting was claimed to be due partly to the shortage of bullock power for land preparation which could extend the rice seedling age in the nursery to more than 25 days. This shortage of bullock power in the villages studies was comfirmed with the estimated land and animal power ratio as it has been discussed previously in section 2.1. The usual peak wet season land preparation activities are in the months of September through December, where water is expected to be sufficiently available from irrigation and rainfall to soften and saturate rice fields to a depth of at least 5 cm. Bernstein (1980) indicated that late commencement of rainfall and insufficient water supply might be responsible for the extended land preparation in con- tiguous areas over several weeks. Water induced delays in land prepara- tion by some farmers can hold up the transplanting of the whole area. He further showed that the implementation of integrated pest control which is built upon the principle of synchronized planting may also require more concentration of labor for land preparation in a certain 18 period. Further, farmers in the study areas explained that the new HYV seedling age was shorter as compared to indigenous varieties. Consequently land preparation should also be done in a shorter time. Thus, the innovation of shortage of HYV, the possible delay of rainfall and water supply, the threat of rice pests and diseases and the shortage of animal power were regarded as land preparation con- straints by farmers in the study areas. CHAPTER 3 AGRICULTURAL DEVELOPMENT THEORY RELATED TO APPROPRIATE TECHNOLOGY AND PREVIOUS RESEARCH IN INDONESIA The literature considered in this chapter is grouped into four categories: 1) agricultural development theory as related to tech- nology, 2) review of rice production programs, 3) experience in tractor- ization in other developing countries of South Asia, 4) previous agri- cultural mechanization research in Indonesia. The agricultural economic development strategy presented will pro— vide some perspective on agricultural development involving small farmers and tractorization. Relevant programs will describe the extent to which mechanization may support rice production systems. The ex- periences of tractorization in the different countries will help identify important factors associated with labor utilization, tractor capabilities, and costs and returns. 3.1 Agricultural Economic Development Theory Related to Technological Changes Schultz (1964) suggested that a significant growth of productivity cannot be brought about by the simple reallocation of existing re- sources in traditional agriculture. Significant opportunities will become available only through changes in technology. He further argued in his book "Transforming Traditional Agriculture," that peasants in traditional agriculture are rational, efficient resource allocators, l9 20 and that they remain poor because there are only limited technical and economic Opportunities to which they respond (in Hayami and Ruttan, 1971). Hayami and Ruttan (1971) suggested that farmers responsetx>changes in production was related to opportunities for more productive technology and improved relative prices. This thesis helped explain the switch from animal power to tractor power in South Asia, where, as Binswanger (1978) pointed out, such change was due primarily to one factor: prices. Hayami and Ruttan expanded Schultz' high payoff model further in their induced development model by incorporating the mechanism by which a society chooses the optimum path of technological change in agriculture. This concept maintains that change in relative price directs the invention or innovation of new and more productive technology. Further, in the induced development model, the mechanisms of induced innovation in private and public sectors interact between technical change and institu— tional development. Dynamic sequences of technical change and economic growth are regarded as critical elements for agricultural and economic development process. McNamara (1972) emphasized that without rapid progress in small holder agriculture throughout the developing world, there was little chance for achieving long term stable economic growth or of significantly reducing the level of absolute poverty (in Stevens, 1975). Stevens (1975) stated that small farmers in developing countries lack capital and were trapped in a technical and economic equilibrium of low productivity and slow growth. Accelerated growth could be generated by transforming traditional farming to more dynamic market 21 oriented farming. The transformation could be enhanced by continuous application of science based agricultural technologies and institutional innovation. Birowo (1977) formulated a strategy for agricultural mechanization development. He suggested four agricultural mechanization related goals for Indonesia: 1) increase labor productivity in the agricultural sector; 2) increase resource allocation efficiency; 3) improve the farm institution, for instance by strengthening the farm c00perative; and 4) facilitate the acceleration of rural industrialization. Gotch (1972) theoretized that the rural growth problem could only be coped with if the broad spectrum of the rural population became a part of modernization. Otherwise the distributive effect of technological change would be of little help. Furthermore, he was concerned about the possibility of technological change producing social and political unrest in the countryside. McLaughlin (1972) and Stevens (1975) concluded that help for small farmers should be through real increase in productivity. Mechanization inputs (including the small tractor) may contribute to the technological change process, but must be adapted to different cultural and economic settings. Overly sophisticated machines are generally inappropriate for the small farmers. Finally, Khan (1977) and Esmay (1978) suggested that local manu- facturing should play a dominant role in the agricultural mechanization development process. Developing nations should not make themselves dependent for the importation of machinery from foreign countries on a long term basis. 22 These selected strategies clearly provide guidance and identify constraints and limitations which are important for establishing any modeling framework for the choice of land preparation techniques. 3.2 Review of Rice Production which Includes Agricultural Mechanization in Indonesia Indonesia began a series of Five Year Development Plans (FYDP) in 1969. The economic development goals were: 1) improvement of rural life by creating more job opportunities; 2) increase of net income per capita; 3) equalization of income distribution; and 4) stabilization of economic growth (Ministry of Agriculture, 1977). Agricultural mechanization projects (including tractorization) have been promoted in conjunction with the BIMAS1 program which is directed toward self-sufficiency of food (rice). The specific objectives of the BIMAS program has been to intensify rice production.2 It has been implemented through: 1. The introduction of new, locally adapted agricultural technology (HYV rice, fertilizers and appropriate pesticides, better irrigation management and improvement of cultivation methods); 2. The facilitation of agricultural inputs at the village unit levels; 3. The providing of low interest credit and subsidiza- tion of agricultural production inputs; l BIMAS , Bimbingan Masa, a kind of extension service model. 2Ministry of Agriculture, 1977. 23 4. The intensification of information dissemination through "siaran pedesaan", i.e., rural broad— casting system, training and visiting, and other extension service means; and 5. The establishment of floor and ceiling prices of rice. The total rice production increase due to this program was 4.6 percent per year during the first FYDP (1969-1973) and 2.5 percent per year during the second FYDP (Ministry of Agriculture, 1977). One of many significant impacts of this program was related to the need for a faster land preparation. Changes in crop management by the introduction of new more productive HYV of rice (IRRI varieties) caused increased labor demand for land preparation during certain shorter periods of the year. Consequently, as these peak labor shortages developed at the village level, farmers began to look for other means to ameliorate this new production constraint. In their induced development model, Hayami and Ruttan (1971) speci- fically identified such an input imbalance or disequilibrium as a dynamic sequence of the development process. It should be regarded as a criti- cal element for the induction of technical and economic growth. Hadisapoetro (1977) concluded that rice production intensification on the existing agricultural land was approaching the point of zero marginal product by 1977-1978. He also predicted that there will be no spectacular discovery of biogenes or agricultural chemicals to increase rice yield in the coming decade. On the other hand, the Ministry of Agriculture estimated that rice self-sufficiency could be achieved by the end of 4th. FYDP (1985) through a 5.6 percent per year increase in 24 rice production. The goal appears impractical without a new and more productively integrated program. Expansion of agricultural land develop- ment outside of Java island through the transmigration program is one alternative. Another promising alternative is the proper selective of agricultural mechanization (Stout, B. A., 1973). Selective agricultural mechanization in this regard is defined as the most appropriate equip- ment or combination of them which can be socially, technically and economically justified. Esmay (1977) learned that Indonesia was in the beginning phase of agricultural mechanization, and proposed the need for a positive selec- tive agricultural mechanization plan in order to improve production and working conditions. 3.3 Tractorization Experiences in Other Developing Countries in South Asia Binswanger (1978) categorized the benefit of tractors in two apparently contradictory views: 1. The substitution view is that a switch from animal power to tractor power was primarily guided by the factors of price and animal power availability. 2. The net contributor view is that power is the primary constraint of agricultural production regardless of tractor prices. The contribution of tractors lends to higher intensification and double cropping, which requires more labor, and results in a higher net return. 25 Esmay (1978) stressed that primary tillage was a critical operation that should be accomplished at the proper time. Chancellor and Singh (1973), in a study on the relationship between farm mechanization and crop yields in a farming district in India found that, in general, there was a slight tendency for higher yield with tractor tillage than with animal tillage, although there were some instances where the reverse was true. Chancellor (1979) recalculated data by Binswanger (1978) and concluded that the mean differences between tractor farms and bullock farms were: 1. Crop intensity + 4.70% 2. Yield per hectare +ll.58% 3. Total crop production per hectare +21.34% 4. Fertilizer and other inputs used +24.35% 5. Labor use per hectare - 2.80% 6. Labor per total production -21.56% McInerney (1973) concluded that the general effects of the intro- duction of tractor technology in Pakistan are: 1) increased farm size, 2) increased crop intensity, 3) caused no benefit changes in crop yield, and 4) increased the total labor per farm but decreased the labor used per unit of cultivated area. Experiments on tractorized as compared to bullock land preparation for rice in South Sulawesi and West Java showed yield increases of 14 percent and 10 percent respectively. Furthermore, hand weeding time was reduced by 20 percent to 40 percent, and the cost of land preparation was 34 percent less for tractor operation (Directorate of Mechanization, 1977-1978). 26 Binswanger (1978) concluded his analysis by stating: l. T‘actors were not responsible for substantially increased intensity, yields, timeliness or gross returns in India, Pakistan and Nepal. 2. Tractors provided the opportunity for land expansion. 3. Tractors shifted the cost advantages of farming toward the larger farms and concentrated land holding to fewer families; 4. The often incredible drudgery of most farm work was not only reduced for the tractor drivers, who usually were the farmer or his son, but also for the rest of the family. As long as there was population growth and slow growth of manufacturing and tertiary sector employment in the rural areas, reducing drudgery was not a social benefit. It was simply redis— tributed benefit from the poorest groups to the already richer strata of rural society. 5. Tractor farms generally did not show much less labor used per hectare than bullock farms. However, this did mean that tractors might not displace agricultural labor. Esmay (1975) found that private ownership of mechanization in Pakistan and Bangladesh tended to displace human labor for crop produc— tion and there was no significant benefit to the small farmers. 27 Chancellor (1970) concluded that almost every farmer hiring tractor service in Malaysia and Thailand, combined the time saved with other resources to produce additional income. The enterprises chosen varied from 65 percent in Thailand to 75 percent in Malaysia that chose enter- prises of agricultural intensification which did not displace labor from the rural to urban areas. In about 80 percent of the cases, income from these new enterprises was sufficient to cover the cost of hiring the tractor. He also found that in Thailand and Malaysia many small work- shops and foundries started making implements for use with tractors and in some cases making power driven puddling machines and small tractors. In 1978, in the Philippines more than 24 companies manufactured IRRI designed machines and in 1974 over 22,000 simple machines were produced by manufacturers and small metal workshops. Nearly half of them had no previous experience in producing agricultural machinery. The references cited demonstrated the contradictory impact of tractorization about which planners should be aware; particularly, if it is viewed from the national development perspective as presented in section 3.2. The other relevant factor connected to this study is that the appropriateness of tractor development is location specific, where it appears to be the dominant factor. This points to the importance of the proper selection of villages for research. 3.4 Previous Agricultural Mechanization Research Work in Indonesia At present IRRI and Rural Dynamic Institute are studying the con- sequences of small rice farm mechanization on production, income and rural employment in selected countries in Asia. Field surveys are being carried out in Pakistan, Indonesia, Thailand and the Philippines. The 28 survey was designed to determine the impact of agricultural machinery at the field, household and village level. The focus is toward the effect of mechanization on production, income and employment. South Sulawesi and West Java were selected for the Indonesia studies (IRRI, 1978). This study is different from the IRRI study in some ways and similar in other respects. The IRRI study consists mainly of a survey on the consequences of mechanization, while this research included more measure- ment of the field performance of technology at the farm level. Both studies are concerned about the effects of farm mechanization on rural societies. Another research study that is relevant to this study was conducted by Gadjah Mada and Bogor universities jointly with the Agricultural Mechanization Division of the Ministry of Agriculture during 1972, 1973, 1974 and 1975. It determined soil draft resistances for land prepara- tion in Java and four other provinces outside of Java, including South Sulawesi. The results helped predict power requirements based upon soil mechanical properties (1972—1975). CHAPTER 4 DATA AND ANALYSIS FRAMEWORK OF FARM SURVEY AND DIRECT MEASUREMENT Data and information regarding choice of land preparation techniques were collected with reference to their relevance to the research objec- tives for which they had been specified as related to the evaluation of tractor performance and its impact on labor use and cost for land prepara— tion at the farm level. There are three sources of data and information: 1) direct measurements of primary tillage operations,20 farm survey on labor inputs for pre-harvest activities, and 3) documentation available at the village and county levels (for supplemental data). This chapter presents: 1) research design, 2) analysis and statistical framework and methodology, 3) data and information, and 4) summary of the statistical estimators. 4.1 Research Desigg A stratified randomized block design was used for the field study. Stratification and blocking provides more precision for statistical estimator and minimizes the undesirable influence of variables unrelated to land preparation and research cost. In addition, the analysis of the data is relatively simple and missing data from individual units can be easily estimated.* The smallest administrative unit, the village, was used in geographical blocking. The population of the study was stratified into three categories: 1) farms using hand labor, 2) farms using bullocks, and 3) farms using tractors. Sampling was performed *Steel and Torrie, 1960. 29 30 by randomly drawing farmers' names from the list of farmers in each village. Geographic, agroclimate, irrigation systems, soil fertility, land use and rice cultivation practices were the criteria used to identify villages for direct measurement of farm tillage operations and survey. A minimum of nine available tractor units was also used as criteria in the selection of villages. 4.1.1 Villages Studied Jatiragas, Jatisari, Rawa Gempol, Bojong Tengah (in Karawang county), Bojong Tengah, Karang Anyar, Mariuk, Tambak Dahan (in Subang county) were selected as the sample villages based upon their demographic appropriateness. Jatiragas and Jatisari villages replaced two proposed villages which had an insufficient number of tractors. Bojong Tengah and Karang Anyar villages of Pusaka Negara district also replaced two previously assigned villages in Pamanukan district because farmers gave inconsistent responses based upon who asked the questions. Demographic characteristics of these eight villages are presented in Table 4.1. Most of the rice land area of these villages was under the Jatiluhur irrigation scheme, which serves a total of 260,000 hectares in the Northern coastal plains of West Java.* The irrigation system guaran- tees the availability of water for double cropping of rice in each year. Irrigation schedule changes due to repair and maintenance of irrigation facilities can cause a delay of the planting date of young rice seedlings. Transplanting time is a critical time for farmers to be accomplished at the appropriate time. Agroclimates identified for these villages were derived from Oldeman (1975). He specified that D-2 zone prevails in Mariuk and Tambak *Bernstein, 1980. 31 .wemc .momeHH> wo cowumw>ounn "mouaom: .Aome ..>uom .uxm .owuwwuooawou moumuoo: m.~ paw moumuoo: Nu ooumEHuwo whoa zufiomamo mxooaasn mam Houomufi .omeHH> osu oucfi wCHEoo mHOHUmHu mvfimuso on sags mama was cowuassmm< .wmeHH> osu a“ mama moan Hmuou wcu Eoum mxuoaazn pom muouomuu ma voumuoao moumuoo: wCHDUmHunam ma momma Hmscme ma poHHfiu cowuuoooua new Howe: o>wuuapoud x Acommom\moumuoocv mufiomamo ">9 moumEHumm ouo3 mxooaaan mam MOuUmHu zn voHHHu mama mo coauuomoumm .omeHH> onu cw coaumanaoa n momeow paw Awumoz omlmav Hoan mama o>wuonpoua HmHOO Ho cones: u an oposs mNooH x Homeoe\m35wme Nun Nun mun mun m m mun mun oomeHHoouwa so .. co mm .. oH N4 mH mooHoH one so 6H OH m oH .. s OH NH .oHooHHso so wH om cm mm OOH ow we ow muuouomuu up wH om om mm ooH ow we oh ARV ooHHHo ocmH Ho ooHouooooa HoHoooooo mm. 8N.H Hm. «H.H ma. oo. as. 06. Away some poo moHoHom mMHH HHHH new mom mwo cmw mNo omq Hmsv ooHu not oooHsoH moNH wwHH Nam omm Hma ONOH Has ANA Hosv HoooH "puma Housuasofiuw< m as OH «a no mm m «a muouomuy HOH HmH an mm mm Hm om Hm oHooHHsm NowH soNH amoH HHm ooHH MHHH aHNH moo NuoooH oHoz "Ho3oa annouasuwuw< H~.o ~o.o mH.o mm.o .m.o oH.H .o.= .o.o HNV nosouw poooH oHoz «H.H mm.o aH.H Ho.o .m.: Ha.H oH.o Hm.o HNV noaouw aoHooHooom mm.HH NH.HH Nq.oH mo.mH Ha.wH HH.oN wN.mH aw.HH NHNV uoomH oHoz ooNoH Hmma mama Nomm memo «Hmw Home aosm Hooooooov Hoooe 1HoaoHv ooHooHoooa o HN o 8 mm an o m HHEHV sasngm me HmH oHH oHH HeH emH mHH wNH Haxv oouoxmw “Eon—w COHquO‘H may a: :.>psum Hops: mommaaw> mo mofiumwuwuomumzo ownmmuonmQ .H.c oHan 32 Dahan, implying three or four consecutive wet months and two to four dry months as the usual pattern of rainfall. A wet month is a month 200 millimeters or more rainfall. A dry month has 100 millimeters or less rainfall. The D—3 zone was the same as the D-2 zone with five to six dry months. The E zone has less than three consecutive wet months and at least five dry months. 4.1.2 Sample Unit and Size Farm households and rice fields were selected as sample units for farm survey and for direct measurement respectively. A rule of thumb used to define sample size is presented in equation 4.1 (Agro- economic Survey Institute, 1977). (n — l)(t - 1) 3_15 (4.1) (n - l)(3 - l) 3_15 n > 9 (rounded) The sample size for each treatment was equal to nine units for each village (either for farm survey or field measurements). Two tillage operations were used with labor and tractor preparation for the rice land preparation. Three tillage passes with bullock Operations were found in Sukatani, Karang Anyar, and Mariuk villages. There was, however, no clear difference between the second and third tillage Opera- tions. A total Of 216 samples were assigned for each farm survey and the first and second tillage measurements. 33 The actual number of samples obtained for this study, was 608 units (see Appendix 1). Survey information pertaining to the full utilization of manual labor was difficult to assess. In Jatiragas, Rawa Gempol and Karang Anyar villages there were only a few farmers who employed only manual labor for both the first and second tillageoperations. These farmers had land that was very muddy and too soft for either trac- tor or bullocks. At the other villages of Jatisari, Sukatani, and Mariuk, few farmers still hired manual laborers, largely because of long standing social relationships. 4.1.3 Statistical Analysis Model The linear additive model of Steel and Torrie (1960) was used as presented in equation 4.2. Xij =11+Ti+Bj +eij (4.2) where Xij = the estimated value Of variable being studied u = the population mean T1 = mean of variance attributed to type of tillage technology Bj = mean of variance attributed to village (blocking) eij = common error developed during measurement survey The basic assumption for the linear additive model analysis of variance, where tests of significance are to be done, is that random com- ponents are independent and normally distributed about a zero mean with a common variance. It is, therefore, necessary to consider whether the 34 (1:... Q | READ DATA COMPUTE x, 52, 8, cv l ? NO NORMAL- LITY 1 FTRANSFORMATION ya 1 S ‘1 l IANALYSIS OF VARIANCE l ? NO SIGNIFI- ANCE YES Y MULTI RANGE REGRESSION COMPARATION ANALYSIS Q J STATISTICAL OUTPUTS ‘J #‘/// SIMULATION I Figure 4.1. Procedure for statistical analysis. 35 population could be adequately described by a normal distribution. Battacharyya and Johnson (1977) suggested a check of normality distributed population with intervals of: u + 10, u - 10 contains .683 u + 20, u - 20 the proba- .954 u + 30, u - 3o bility .997 where probability = percent of the population The number of Observations outside the symetric interval about the means of normally distributed population can be counted and divided by the sample size "n". The result gives the relative frequency which is com— pared with theoretical probability Of roughly 1/3, 1/20, an 1/300 (11). The expression for normality estimation is: r - Pl > 3 (4.3) VP(1-P)/n where Observed relative frequency distribution outside '0) II the symetric interval about the means p = theoretical probability of frequency distribution outside the symetric interval about the means n = number of samples, which is set to be 9 (in this research) If the left hand side of this inequality is greater than 3, it would indicate lack of normality and in this case data needs to be transformed 36 (either square root or logarithm), before continuing further statis- tical analysis. In this research the sample size "n" was nine, thus the number of observations outside the symetric interval should not have exceeded four. The normality test indicated that plot size data of the first tillage Operation needed to be transformed before ANOVA analysis was carried out. The computation of statistical estimators for the expected means (u), variance (02), standard deviation (0) and coefficient of variation (cv) are presented in equation 4.4, 4.5, 4.6 and 4.7 respectively (49). 2 2 (Exij) S = (Exij2 - ——n— )/(n - 1) (4.5) s = /52 (4.6) 0v = % x 1007. (4.7) where 'R = an estimate of population means (u) 52 = sample variance, an estimate for population variance (02) S = standard deviation, an estimate for 0 CV = coefficient of variation The analysis of variance (ANOVA) is essentially an analysis of the error mean square (02), an estimate of common error. The expected error mean square (calculated F-value) in ANOVA is defined as a ratio of the independent estimate of the same population variance (02). Multiway classification, as suggested by Steel and Torrie (1960) which deals with two or more criteria, was used for this ANOVA. 37 Table 4.2. Analysis of variance for choice of land preparation. Source df MS EMS(F) . 2 2 Village (v-l) = 7 0 +30 E V 2 2 Technology (t—l) = 2 0 +80 5 t 02 +802 6 t 2 o2 Error (v-l)(t-l) = 14 O€+ E Total vt - 1 = 23 The variance mean square: of village effect, in which it contains 2 . . treatment and common measurement errors (O+t.Ov) 1s computed by apply1ng equation 4.8; similarly for residual or common error (02) and variance mean square of treatments (O:+v.oi) by using equation 4.9 and 4.10 respectively. 03 = X.§/t-C (4.8) 02 = X. -C (4.9) t 1./v o: = TS-SE-Sg (4.10) where xij = individual Observed data X.j = village sum of square Xi. = technology sum of square v,t = number of village and treatment respectively C = correction factor, which is computed by equation 4.11 38 X... = grand total (sum of all data) c=x /v.t (4.11) If analysis of variance indicates that thereyun;a significant difference between treatment means (manual labor, bullock and tractor performance), then multi-range comparison using the LSD (least square difference) method and regression analysis were performed. Linear regres- sion analysisrwnscarried out to study the correlation between: 1) plot size and time for land preparation, 2) plot size and residual untilled land, and 3) field capacity and soil draft resistance. The LSD method is basically a student £_test using pooled error variance. For instance, if the difference between bullock and tractor capacity means is to be significant at the confident level, then this difference must exceed the LSD value which is calculated as follows (49): LSDa = tonso (4.12) S— = “2.3% “'13) C where to = tabulated £_value at a given df (degree of freedom) and confident limit (a), in this case, tabulated £_(df = 7,cx= 5%) = 2.365 (two-tailed) S__= standard difference between treatment means 0 In this present research the number of villages (v) was eight, therefore: S = 0.5 MS? (4.14) E 39 4.2 Relevant Criteria for Land Preparation Measurements Criteria to determine relevant data related to land preparation techniques were derived from Hunt (1973). He suggested that the soil tillage Operation can be evaluated through the analysis of capacity which is a measure of work done relative to time (rate of work). The mathematical expression of this concept is presented in equation 4.15 (Hunt, 1973) PC = M13041); (4.15) where PC = field capacity (hectares/hour) W = width of cut (centimeters) S = traveling speed (kilometers/hour) E = field efficiency (%) 10 = conversion factor to a unit hectares/hour Width of cut was assumed equal to the width of implement attached to the land preparation equipment: moldboard plow and comb harrow for bullocks, hoe for manual labor, or rotary tillers for trac- tors. Field efficiency was defined as the time Of actual field work divided by the total time spent in the field; including idle, minor adjustment, refuel and other uneffective Operation time. The machine or equipment efficiency was defined as the theoretical capacity divided by the actual time to complete tillage Operation for one hectare of rice field. Other technical criteria for the evaluation of land preparation techniques are elements related to power requirements such as: depth 40 of tillage, plot size, mechanical properties of soil, tractor size in terms of horsepower (hp), water and soil condition, working environment such as degree of temperature, relative humidity, and operator skill and age. 4.2.1 Age and Experience of Land Preparation Operators Age and experience of Operators were two factors that presumably influence the degree of field capacity. Data related to these factors were obtained from interviews conducted in the rice field and presented in Table 4.3 and 4.4. Table 4.3. Average ages of land preparation Operators (years). Village Manual labor Bullock Tractor Jatiragas 37.5 36.0 29.8 Jatisari 37.9 34.4 30.8 Rawa Gempol 36.0 31.4 30.1 Sukatani 32.2 32.4 28.1 Bojong Tengah 39.4 34.5 30.3 Karang Anyar 34.7 35.8 28.9 Mariuk 38.3 35.9 29.6 Tambak Dahan 47.6 43.3 26.2 Average: 38.8 36.6 29.2 Standard Deviation: 5.2 4.4 1.5 Coef. Variation (%): 13% 12% 5% Table 4.3 suggests that manual labor and bullock operator ages didrun:differsignificantly (5% level). On the other hand, tractor 41 operator agesynnxzsignificantly younger than either manual laborers or bullock operators (1% level). The age range, as listed in Appendix 3 , served as one factor to estimate the potential number Of male manual labor participating in land preparation by village. Table 4.4. Operator experience in land preparation (years). Village Manual labor Bullock Tractor Jatiragas 16.5 15.5 2.0 Jatisari 17.3 14.2 1.5 Rawa Gempol 20.0 13.9 1.2 Sukatani 15.1 14.4 2.4 Bojong Tengah 23.8 14.8 1.1 Karang Anyar 17.8 19.5 1.8 Mariuk 24.3 25.6 1.7 Tambak Dahan 27.0 24.3 0.9 Average: 20.2 17.8 1.6 Standard deviation: 4.3 4.8 0.5 Coef. Variation (%): 21% 27% 31% Years of Operator experience are presented in Table 4.4 and Appendix 4 . The average yeansof experience for tractor operators was 1.6 years as compared with 18 and 20 years for bullock and manual labor experiences respectively. The tractor technology in Tambak Dahan was in the introductory stage with an average:operator experience of less than one year. The longest tractor Operator experience was found in Sukatani which was 2.4 years. 42 4.2.2 Hours Worked Per Day The range Of temperature and relative humidity in rice fields in the course of the day time were 22-31 centigrade and 89-91 percent respec- tively.* It was common in the study areas for three to four operators to work one tractor in shifts, while for bullock Operation there was only one operator. Therefore as indicated in Tables 4.5 and 4.6, tractors were operated for more hours per day. Table 4.5. Hours worked per day on the first tillage operation. Village Manual labor Bullock Tractor Eff. Idle Eff. Idle Eff. Idle Jatiragas 7.7 1.7 5.2 0.5 15.4 1.8 Jatisari 6.8 1.2 5.2 0.4 13.3 1.8 Rawa Gempol 7.7 1.3 6.7 1.0 10.0 1.6 Sukatani 6.6 1.2 5.1 0.4 13.3 1.4 Bojong Tengah 7.4 1.5 6.9 0.9 9.6 1.1 Karang Anyar 8.3 1.5 6.1 0.8 11.7 1.9 Mariuk 8.0 1.5 5.1 1.0 12.8 2.5 Tambak Dahan 7.9 1.8 7.0 1.4 10.1 1.5 Average: 7.5 1.5 5.9 0.8 12.0 1.7 Standard Dev.: 0.6 0.2 0.8 0.3 2.0 0.4 Coef. Variation: 8% 13% 14% 37% 17% 23% TVi VVV V vv ‘ w v Ti * Directorate of Geophysic and meterology, 1975. 43 Table 4.6. Hours worked per day on the second pass of tillage Operation (hours). Village Manual labor Bullock Tractor Eff. Idle Eff. Idle Eff. Idle Jatiragas 6.3 0.9 4.7 0.3 12.8 2.3 Jatisari 6.8 1.3 5.7 0.5 9.3 0.6 Rawa Gempol 7.7 1.5 4.9 0.4 9.6 1.4 Sukatani 7.7 1.3 n.a. n.a. 15.0 2.2 Bojong Tengah 8.2 1.7 7.1 1.6 10.3 1.6 Karang Anyar 8.9 1.9 6.9 1.4 11.4 2.1 Mariuk 7.7 1.8 5.6 1.0 10.0 2.6 Tambak Dahan 7.6 0.4 5.2 1.7 11.4 1.9 Average: 7.6 1.3 5.7 1.0 11.2 1.8 St. Deviation: 0.8 0.5 0.9 0.6 1.9 0.6 Coef. Variation: 10% 37% 16% 60% 17% 33% Note: n.a. = data were not available. If necessary, tractors may be operated almost 24 hours per day, as was found in Jatiragas (see Appendix 5 ). Draftanimahsparticularly water buffalo (bullocks), which have dark skin, cannot work for more than three hours consecutively. They need to be cooled down by allowing them to submerge in water for 15 to 30 minutes or by splashing them with water. The actual time used for the first and second field operation wasdefinedas effective time. Idle time included the time for rest, eating, minor adjustment and repair, refueling and resting for draft animals in the field. The number of working hours per daytnnsan ‘ [i.(/ V ls , y ,_11 l J six square centimeters and the cone angle was 30°. Measurement was r" A .’ 44 important measure of the seasonal capacity for land preparation. WOrk- ing hour data were recorded based upon one full day of Operation. For some verification the operators were asked whether they had worked normally during the recording day. If, for instance, they had not, they were asked for any possible deviations. This check was particularly important for tractors as they are mobile. Sometimes the tractor opera— tions started as early as 5 o'clock in the morning and left as late as 1 o'clock after midnight or sometimes tractors only worked a few hours before moving to another village. A highly significant greater number of hours worked were found for the first and second operations of trac- tor technology as compared to either bullock and manual labor (1% level). The idle time was not significantly different between tractor and hand labor in either the first and second operations, however, bullock idle time was significantly higher relative to the tractor for the first and second operations (5% level). 4.2.3 Depth of Tillage Depth of tillage and soil mechanical properties as measured in kilograms/square centimeter are two factors associated with power re— quirements for land preparation. Data regarding these measurements is presented in Table 4.7. Depth of tillage was measured only for the first operation. The softening process of the top soil in the second opera- tion could not be differentiated as to whether it was caused by second tillage Operation or RX1K§ES§f This situation made the depth measurement for the depth of the second Operation to be not aggggate. A spring type N {icone penetrometer was used to measure soil hardness. The cone area was ‘\\H____' _____ ‘ / T~'/ I“ 45 Table 4.7. Depth Of tillage and soil mechanical properties. Depth of tillage Soil mechanical properties Village MLl BL2 TR3 Hardness PI (cm) (cm) (cm) (kg/cu?) <7.) Jatiragas 6.08 12.48 10.71 1.762 36.60 Jatisari 6.41 11.30 9.92 1.583 35.78 Rawa Gempol 6.41 10.79 10.25 0.825 35.63 Sukatani 5.60 11.23 8.80 1.300 33.69 Bojong Tengah 6.10 10.92 10.50 1.297 35.84 Karang Anyar 6.16 10.13 11.29 1.243 34.15 Mariuk 7.45 9.97 11.23 0.917 29.06 Tambak Dahan 5.30 9.08 10.06 0.726 29.42 Average: 6.19 10.71 10.34 Std. Deviation: 0.64 1.05 0.80 Coef. Var.: 10% 10% 8% lManual labor 2Bullock 3Tractor taken from the surface down to 40 centimeters. Figure 4.1 illustrates the measured gradient of soil hardness profiles. The index of soil plasticity (PI) associated with soil hardness was also considered as an essential element for evaluation of tractor performance for land prepara- tion.* A total of 81 soil samples for depths Of 0, 10 and 20 centi- meters were taken and analyzed at the Gadjahmada University soil laboratory to determine the soil plasticity index. *Kishu, 1972. a a. 0| 0 III -- Depth cm -- N 0 d 0 0| -- Depth cm -- u 20 25 46 . O \ \ \ I \ -Jat1ragas \\ " O \ \ \ \ \ \\ ‘\ \ \ \ \ _ \ o \\ \ \ \ \ O p .u I I J 1 I Q J 0 1 2 3 4 5 -- Hardness kg/cm2 -- p o \ \ \ \ \ \ - % -Rawa Gempol \ \ \ \ D O \ \ \ \ D O \ \ \ \\ n ‘0 \\ \ \ \ 1 1 4‘1 1 j o 1 2 3 4 5 Hardness kg/cn2 -- Figure 4.2a. -—— Depth cm -- -- Depth cm -- M d 6 d In 20 25 U! .5 O u‘ M N O 25 Soil hardness " o \ \\ \ -Jatisari \ \ I- o I I I I .- 0‘ ‘\ ‘\ - O \ \ \ I. O I I I _I l I j o L _J I 1 2 3 4 5 -- Hardness kg/cm2 -- " O \ \ \ \ \ . . K -Sukatan1 \ \ \ \ - O ‘0 ‘ O I. 0 \ \ \ \ J I A1 J 4 o 1 2 3 4 —- Hardness kg/cm2 -- profiles. 47 0-0\ 0-0 \ \ I \\ ‘\ I . I ‘\ -Bogong Tengah I X ‘— Karang An ar 5- I \ Y E o‘ I 5- o \ \ \ \ \\ E ‘\ 5'°' ‘0 '5 10- 0 o. \ m \ Q) \ m \ Q \ Q \ ' 15 - \ ' \ I o l 15- o I ‘ ' I \ \ I \ I \ I 20' q zo- 3: ‘ \ \ \ \\ \ 25 I 4 0| L 4 25 I I \o I I I 0 1 2 3 4 5 O I 2 3 4 5 -- Hardness kg/cm2 -- -- Hardness kg/cm2 -- 0-9 0'9 I‘ \ ‘ \ I . \ I ‘ I I ' 5 - o : s- Q 5 X - Tambak Dahan \ - Mariuk I‘ E ‘ \ I .c .8 ‘ 10’ o u 10- ‘ 19} K 8 9. o \ o \ I ‘ | \ '15- 0 I15. 0 I " I \ I \\ I \ ‘ \ \ \ \\ I‘ \ 'I 25 ' 0 . L I , 25 444 I o -1 I I; 0 I 2 3 4 5 O I 2 3 4 5 -- Hardness kg/cm2 —- -- Hardness kg/cr.2 -— Figure 4.2b. Soil hardness profiles. 48 4.2.4 Plot Size and Percentage of Untilled Land It is generally hypothesized that plot size effects land prepara- tion equipment field capacity.* Therefore, plot size (dimensions) was measured as well as time for tillage operation and the residual of un- tilled land on the corners and levee sides of rice field. Manual labor was used to complete land preparation after bullock and tractor opera- tions. There was always some residual untilled land after bullock and tractor operations. Large steel wheels attached to tractors, to avoid sinkage, prevented tractors from approaching the plot corners and levee sides closely and resulted the increase of untilled areas. Similarly, long wooden plow bars used in bullock Operations caused difficulty in maneuvering the draft animal around the corners of the rice fields and resulted in untilled land on those corners and levee sides. Table 4.8 shows that plot size tilled by tractors was significantly larger than for either bullocks or manual labor (5 percent level). Table 4.8. Plot size and residual untilled rice land. , Plot size (ha)* Residual land (%)** Village HL BL TR BL TR Jatiragas 0.165 0.107 0.147 7.61 7.61 Jatisari 0.103 0.123 0.158 4.09 8.38 Rawa Gempol 0.889 0.130 0.167 3.43 6.43 Sukatani 0.126 0.169 0.154 4.59 6.89 Bojong Tengah 0.144 0.144 0.150 2.86 8.10 Karang Anyar 0.108 0.116 0.166 3.81 6.78 Mariuk 0.165 0.138 0.209 3.96 7.78 Tambak Dahan 0.126 0.088 0.149 4.38 6.11 Average: 0.128 0.127 0.163 4.34 7.26 Standard Deviation: 0.028 0.024 0.020 1.43 0.82 Coefficient Variation: 22% 19% 12% 33% 11% *Average of first and second operation. **First operation only. *Hunt, 1973. 49 The 7.26 percent of residual untilled land left by tractors was sig- nificantly higher (1 percent limit) than the 4.34 percent left by bullock tillage. A linear regression indicated a weak correlation between plot size variation and percentage of untilled residual land, with a coeffi- cient of determination of 0.08 and 0.01 respectively. This implied that for a plot size range of 0.09 to 0.17 hectares, the percentage of untilled land was quite independent from plot size. 4.2.5 Travelling Speed of Bullocks and Tractors The traveling speed is associated with field capacity and power requirement for the tillage operation. The rate of traveling speed was recorded in seconds for distances of either 10 to 20 meters of tillage operation. Tractors were found to operate nearly as fast as bullocks (see Table 4.9). An LSD analysis indicated that the difference of travelling speed was highly significant (1 percent level). Bullocks pulled a plow (18-25 cm wide) for the first operation and a comb harrow (125-160 cm wide) for the second and third passes, while rotary tillers with 20 to 24 blades (54-64 cm wide) were used by tractors. 4.2.6 The Actual Field Capacity The total time required to complete the land preparation was measured for each rice plot. The total time consisted of the net (effective) time plus the idle time. Conversion of measured time in minutes per plot to hours per hectare were made and the results are presented in Tables 4.10 and 4.11 for the first and second operations respectively. 50 Table 4.9. Travelling speed of bullocks and tractors (km/hr). Village First operation Second operation Bullocks Tractors Bullocks Tractors Jatiragas 1.33 2.94 1.22 2.03 Jatisari 1.36 3.28 1.01 2.64 Rawa Gempol 1.19 2.31 1.11 2.42 Sukatani 1.25 2.03 1.03 2.06 Bojong Tengah 1.30 1.81 1.01 2.44 Karang Anyar 1.14 2.31 1.33 3.03 Mariuk 1.14 2.25 1.01 2.14 Tambak Dahan 1.08 2.01 1.33 1.92 Average: 1.22 2.37 1.13 2.33 Standard Deviation: 0.10 0.50 0.14 0.37 N N I—‘ N H L» N I—I C‘ N Coefficient Variation: 8 In summary of Tables 4.10 and 4.11, the average effective time re- quired for tillage in hours was 309.6, 52.3 and 14.8 per hectare, for two passes for manual labor, bullocks and tractors respectively. A slight negative regression slope along with a small coefficient of determination, of less than 0.1, suggested that within a plot range of 0.09 to 0.17 hectares there was a weak association at 5 percent confident level between plot size and time efficiency. 4.2.7 Fuel and Oil Consumption Diesel fuel consumption was recorded based upon operating time, rather than unit areas. This approach was considered more adequate, as tractor operators usually filled the fuel tank before they started in 51 Table 4.10. Average time required for the first tillage of one hectare of rice land (hours). Village Manual labor Bullock Tractor Jatiragas 241.8 40.7 6.3 Jatisari 240.0 31.3 8.3 Rawa Gempol 134.0 28.0 8.3 Sukatani 149.3 36.2 8.0 Bojong Tengah 134.0 31.8 11.3 Karang Anyar 140.7 36.8 7.7 Mariuk 240.0 48.3 8.2 Tambak Dahan 247.2 47.8 10.8 Average: 190.9 37.5 8.6 Standard Deviation: 55.2 7.5 1.6 Coefficient Variation: 29% 20% 19% the morning and again at lunch, dinner or rest time. Measurement was made by marking a fuel stick when full. The fuel level was exactly filled to this mark each time the measurements were made with the tractor standing exactly horizontal. Data regarding fuel and oil consumption are presented in Table 4.12. The amount of fuel used was recorded in cubic centimeters/minute, then converted to liters/hour. Oil consumption was not measured directly. Tractor operators were asked to provide information on oil use and this was checked by physical observations. Operator information was generally confirmed as being accurate. Table 4.11. Average time required for the second tillage of one hectare of rice land (hours). Village Manual labor Bullock Tractor Jatiragas 126.0 20.7 6.7 Jatisari 79.0 15.8 5.2 Rawa Gempol 98.3 15.0 5.7 Sukatani 112.3 29.5* 6.7 Bojong Tengah 122.2 11.7 5.8 Karang Anyar 79.5 27.8* 5.7 Mariuk 143.5 27.0* 6.8 Tambak Dahan 189.2 13.3 7.3 Average: 118.7 14.8** 6.2 Standard Deviation: 36.3 2.7 0.7 Coefficient Variation: 30% 18% 12% *Second and third passes. **Average for second pass only. The higher fuel consumption rate in Jatisari for both the first and second operations coincided with higher soil hardness values and the higher travelling speeds (see Tables 4.7 and 4.9). In Tambak Dahan, lower fuel consumption for the first operation was found along with the lower degree of soil hardness and slower travelling speed. Also, most tractors used in this village were relatively new (see Tables 4.4, 4.7, and 4.9). In general, fuel consumption levels varied only slightly between villages. The coefficient of variation was less than 15 percent. The average oil consumption was found to be 0.042 Table 4.12. Average fuel and oil consumption for tillage operations (liter/hour). Village First pass Second pass Oil Jatiragas 1.249 1.183 .047 Jatisari 1.486 1.330 .036 Rawa Gempol 1.192 1.122 .036 Sukatani 1.276 1.174 .049 Bojong Tengah 1.030 1.067 .030 Karang Anyar 1.308 1.109 .040 Mariuk 1.199 1.285 .050 Tambak Dahan 0.968 1.171 .050 Average: 1.214 1.180 .042 Standard Deviation: 0.162 0.088 .008 Coefficient Variation:13% 7 liters per hour. This level of consumption for 7-8.5 horsepower trac- tors was close to the level recorded by Mechanization, which was 0.8 liters per 4.2.8 Locally Made Tractor Performance Field measurements of locally made The Directorate of Agricultural 100 horsepower-hours. tractors (7-8.5 hp) were con- ducted in Gabus Wetan village, located about 90 kilometers from the nearest sample village. the data are presented in Table 4.13. type) tractor was powered by a diesel engine. Sprockets, chains, Five tractor measurements were conducted and This locally manufactured (IRRI pulleys and two v-belts were used for transmitting mechanical power Table 4.13. 54 Field performance data on locally made tractors. Imported* Plowing Harrow (average) Depth of cut cm 10.0 -- 10.34 Traveling km/hr 4.8 4.3 2.33 Width of cut cm 24.3 125-160 64 Fuel consumption 1t/hr 0.941 1.327 1.214 Oil consumption lt/hr 0.034 0.042 Field capacity hr/ha 12.5 3.83 8.6 Machine efficiency % 56 37 62 Plot size ha 0.05 0.115 0.167 Residual land % 12 -— 7.26 Worked hours/day a. mean hr 9.7 -- 12 b. std. dev. hr 0.94 -- 2 Price Rp. 1,000 1,650 -- 2,330 Operator wage % 15 -- 14.3 *The average imported tractor data are provided for comparison. These are first operation data. from the engine to the steel wheels. probably in part due to the lack of steering clutch. The lower plowing efficiency was A new design, with steering clutch was still being tested when this research was conducted so was not yet available for field use. 55 4.3 Labor Use and Cost for Land Preparation Derived from Farm Survey The level of agricultural supply, price level, inflation and interest rate and their complex linkages were assumed as exogenous factors to the farmers. Farm survey, Operator interviews and litera- ture studies were conducted to obtain relevant exogenous data. Land- holding of the sample farms, rice yield and production inputs were also recorded to find their association with cost and labor use for land preparation. 4.3.1 Equipment Price and Operator Wages Price levels for hoe, bullocks and tractors are presented in Table 4.14. Tractor price was based upon the price of the "standard" unit of small tractors for rice fields, which include floating steel wheel, rotary hoe (blade) set and rubber tires. Higher tractor prices were found in Jatiragas, Rawa Gempol and Karang Anyar and coincided with greater proportion of land and bullocks in those villages (see Table 4.1). While the highest tractor price was found in Tambak Dahan, Sukatani and Mariuk had the lowest tractor prices and also experienced a longer time of tractorization. In Jatisari many sample farms obtained their lower tractor price by direct purchase from the dealers rather than through the Rural Bank. The bullock price was found lowest in Mariuk, possibly because this village has the highest number of either bullocks or tractors. Most tractors in the sample villages (except in Sukatani, Mariuk and Jatisari) were obtained on credit through the Rural Bank at 10.5 percent to 12 percent interest rate yearly, with six seasonal (six 57 Table 4.15. Operator wages. Village Mangldigbor 3:11)}3231 Tragtorz Jatiragas 538.5 1656.2 15.0 Jatisari 558.3 1533.3 15.0 Rawa Gempol 513.3 1366.7 15.0 Sukatani 542.9 1280.0 11.6 Bojong Tengah 508.8 1933.3 16.3 Karang Anyar 494.1 1256.5 14.3 Mariuk 505.9 1142.8 16.7 Tambak Dahan 500.0 2142.8 10.6 Average: 520.2 1538.9 14.3 Standard Deviation: 23.2 352.0 2.1 Coefficient Variation: 4% 23% 15% 1Including cost for bullocks. 2Percentage of custom-rate. Lower operator wages for all technologies were found in Karang Anyar. The highest wage rate for bullock operations was found in Tam- bak Dahan, where worked hours per day for this technique was the longest (first operation, see Table 4.4 and 4.5). Lower bullock operator wages in Mariuk and Sukatani, along with low manual labor wages in Mariuk coincided with the high tractor population in those villages (see Table 4.1). 58 4.3.3 Land Holdinggof Sample Farms Land holding in Northern plains of West Java, like in other parts of the country, generally classify the economic status of farmers. The average land holdings of sample farms are presented in Table 4.16. Table 4.16. Land holding by sample farms (hectare). Farm utilizing: Village Manual labor Bullocks Tractors Jatiragas 0.171 0.62 0.89 Jatisari 2.03 1.15 2.21 Rawa Gempol 0.92 3.70 2.30 Sukatani 1.05 1.73 2.90 Bojong Tengah 1.13 2.05 1.01 Karang Anyar 1.28 1.94 5.062 Mariuk 1.82 1.26 4.30 Tambak Dahan 1.34 1.05 1.94 Average: 1.22 1.69 2.58 Standard Deviation: 0.57 0.95 1.47 Coefficient Variation: 46% 56% 57% 1Four samples. 2One sample farm of 15 hectares. The average land holding for farms using tractors was significantly larger as compared to those using manual labor or bullocks (1 percent level). Furthermore, the larger tractor farm land holdings in Sukatani, Rawa Gempol and Mariuk, coincided with the greater number of tractors in those villages. 59 4.3.4 Labor Inputs for Preharvest Rice Production Activities The proportion of labor input for land preparation as compared to the total preharvest labor input in the rice production system is im- portant. The numeric values for labor input as given by the farmers, are shown in Table 4.17. Table 4.17. Labor input for preharvest rice production activities. Village gital ingit (hr)TR LanngrepgiationTé%)* Jatiragas 1106 555 432 39 21 13 Jatisari 1150 696 697 45 41 35 Rawa Gempol 1273 678 625 48 26 24 Sukatani 1241 808 809 36 36 25 Bojong Tengah 1699 1110 946 38 22 18 Karang Anyar 1209 771 858 37 24 19 Mariuk 1342 871 745 47 37 29 Tambak Dahan 1175 747 817 60 34 34 Average: 1274 779 741 44 30 25 Std. Dev.: 202 163 159 8 8 8 Coef. Variation: 16% 18% 21% 26% 21% 31% Note: Figures are rounded. *Percentage of the total preharvest rice production labor input. Table 4.17 shows that the labor input on farms using manual labor is significantly greater than on bullock and tractorized rice farms. There appears to be no significant difference in labor input between bullock and tractor farms. This table also indicates that total preharvest 60 labor input on bullock and tractor farms was not significantly different (5 percent level). The percentage of labor for land preparation rela- tive to the total preharvest labor input shows highly significant differences between manual labor, bullock and tractor farms. 4.3.5 Fertilizers and Chemicals Used in the Sample Farms Fertilizers commonly used in the study areas were urea, and triple or double superphosphates. The fertilizer doses for each were recom- mended by the BIMAS program to be around 200 to 300 kilograms of urea and 100 to 150 kilograms of superphosphate per hectare. Other agricul- tural chemicals used for controlling rats, stemborers and brown hoppers are diazinon, sumithion and phosphorus. The amounts of chemicals applied are presented in Table 4.18. This table indicates that the average application rate of fertilizers and chemicals does not differ significantly between farms using manual labor, bullocks or tractors. The application of these chemicals was highest in Karang Anyar and lowest in Jatiragas. Most of the rice varieties used were IRRI types (IR-26, IR-28 and IR-36). The occurrence of more frequent rice damage by pests and diseases in Karang Anyar might cause farmers to plant more rice seed to provide some reserve. Rawa Gempol used the least amount of seed. This village receives more sunshine (agroclimate zone E) and has better crop manage- ment along with highest yield per hectare. Table 4.19 demonstrates that farms using tractors obtained a significantly higher yield (5 percent level). 61 Table 4.18. The amount of fertilizers and chemicals used in sampled rice farms (per ha). Village Miertiligirs (kg;R Mghemicafii (lite;:) Jatiragas 414 296 337 2.9 1.6 2.0 Jatisari 338 357 374 4.5 1.8 2.6 Rawa Gempol 345 357 316 3.7 4.5 2.4 Sukatani 286 343 286 3.6 5.6 3.6 Bojong Tengah 356 319 357 4.6 6.3 5.6 Karang Anyar 333 360 429 6.7 6.4 9.0 Mariuk 375 325 331 5.4 3.5 4.9 Tambak Dahan 337 335 343 3.8 3.2 4.5 Average: 348 333 347 4.4 4.1 4.3 Standard Deviation: 37 30 42 1.2 1.9 2.2 Coef. Variation: 11% 9% 12% 27% 46% 53% Table 4.19. farms. 62 Level of seeds planted and rice yield on the sample Village Seeds (kg/ha) Yield (ton/ha)* HL BL TR HL BL TR Jatiragas .a. 29 27 4.1 .5 4.9 Jatisari 26 31 23 4.4 .5 4.6 Rawa Gempol 31 29 20 3.8 .8 6.8 Sukatani 29 28 27 4.6 .6 4.3 Bojong Tengah 24 25 23 3.9 .7 4.7 Karang Anyar 43 34 34 4.0 .2 4.5 Mariuk 29 31 26 4.0 .1 4.0 Tambak Dahan 31 27 27 3.9 .5 5.0 Average: 30 29 26 4.1 .5 4.8 Standard Deviation: 6 3 4 0.3 .2 0.8 Coefficient Variation: 21% 9 16% 7% 17% *At the harvest moisture content. CHAPTER 5 DEVELOPMENT MODEL AND SIMULATION FOR LAND PREPARATION TECHNIQUES Perhaps it is typical that developing countries, such as Indonesia, have relatively frequent, complex and uncertain problems that need immediate relevant information for decision makers to use in planning. The choice of land preparation techniques for rice cultivation is a case in point. Traditional analyses are not sufficient to deal with the problems of availability of agricultural power at peak seasons; commencement of rainfall as related to suitable time for tillage opera— tion; or the control of the outbreak of rice pests and diseases (e.g., rat, stemborer, brown hopper). A computer assisted systems analysis approach can serve as an effective means for dealing with such problems. This chapter presents a computer aided systems analysis approach for prediction and comparative analysis for alternate choices of land preparation techniques in the Northern coastal plain of West Java. While researchers traditionally have relied on conventional, large computers for such analyses, they are not always readily available in many developing countries. For this reason, a micro computer with BASIC language was used for analysis and simulation of the choice of land preparation techniques in the present research. 63 64 5.1 A Systems Approach for Comparing Preparation Alternatives A system is defined as a set of elements or components which interact or link with each other in the performance of given functions (Manetsch g£_gl., 1974 and Naylor g£_al,, 1968). Components are differ- entiated into: 1) exogenous (environmental) variables, 2) fixed or controllable input (endogenous) variables, 3) input parameters, and 4) output variables. An imaginary line or systems boundary separates endogenous systems components from exogenous (environmental) variables. While exogenous components influence the system's behavior or performance, conversely the system itself does not have, or has only a weak influence on the exogenous variables. Manetsch g£_§1, (1979) suggested four major phases in the applica- tion of the system approach process: 1) feasibility evaluation, 2) abstract modeling, 3) implementation design, and 4) system operation and implementation. Simulation is associated with the last three phases. 5.1.1 Identification of System Components The first step is feasibility evaluation, which includes the process of identifying the system components. The need for analysis, as well as components, linkages and functions of land preparation system are identified and illustrated in Figures 5.1 and 5.2. Weather, irrigation schedules, prices, population, rate of inflation, non-agricultural em- ployment and rice pests and diseases are all defined as exogenous com- ponents. Input parameters, as system design elements, serve to specify the structure of the system. These parameters tend to be fixed and are 65 .mao>oa omwaaw> cam Ehwm pm moswquoop coapwpwmoem Unwa mo mowono you awemwww xopMome .H.m opswwm Moan wwmwmpqu pmoo posmwm mawpcwag mo hwaon mapwnfimmcna cowposconm ommmmnonH Ampfiawdu mmwafifip nmppmmv nOHPwNfiHfips momma adefipmo mmmaHHwEwp mpmfipmopmm< owpwn oedema :80 SSE mapwpfimmo I 38883 pnmapmsmo wfinopwpo pqoagoam>o© Honowpwc .amoa cam pmohmpnfi Boa puma deem .:0Hpmxwp .aofipwuwcfimpfim Ahopwav . A>wmv cmmm . mmmem2¢moH anwm one no nofipw>fipaso moan you fIIIIII. mesmzH mamaqqomazoo wannabe I mosvfiqnomp soapwnwmohm @cwa mo mafiono posom nonowse hmsom Hmsfiq< yonwa stnmz 96309 was 0 wepmzH mu omezooz: H pa: “pm< petJBA eq 0; momwomfic cam mpmog wows .pathoagEo HandeSOHnmwao: .cowpwamafi .zowpwa Ismom .mowpm .madconom nowpwmflpmfi .nmnpwo; umbozmooxm 66 .Hm>oH Ehmm osu um moscficnomu coaumumawua puma mo mofiosu pom Hobos maooH.222:2ovoHMfiHaEwm .N.m muawam .cofiuomumucfl o> m mmfio "IIIII muHmm MUHAOm a» u U .zmm>ow I J \memmm "~on ZOHamq IDmOm >m I INPUTS, READ , INITIALIZATION AVAILABLE, SUITABLE AND EFFECTIVE TIME ANALYSIS L SYSTEM CAPACITY 1 SELECT COMPUTE LEAST COST EXPENSES YES COMPUTE RETURS COMPUTE WAITING DAYS YES I! PROFITABILITY LABOR USAGE AND INDEX TOTAL COST SIMULATION OUTPUTS '- UP-DATING 6th SEASON YES (:‘S T O P > Figure 5.3. Simplified flow chart for land preparation model at the farm level. No ' I 72 Q-values are obtained through a computer random number generation program. The values for u and O are derived from statistical inferences. For computer programming convenience, equation 5.1 was modified as: X = O x (Q - 12/2) / 712/12 + u (5.2) In this case the value of M (number of iteration or looping) for com- puter program is set equal to 12. Thus equation 5.2 becomes: X = O x (Q - 6) + u (5.3) The computer subroutine to find Q from a normal distributed random probability is designed as follows: 4980 ****** SUBROUTINE RANDOM PROBABILITY ****** 4990 S = 0.0 5000 FOR K = 1 TO 12 5010 A = RND (l): S = S + A 5020 NEXT K 5030 Q = S - 6 5040 RETURN 5050 END The other extensively used subroutine is the waiting time (in days) for predicting availability Of hired manual labor or animal power as described in the following computer program: 73 4840 ****** SUBROUTINE WAITING TIME FOR RAINFALL ****** 4850 ****** AND AVAILABILITY OF TRADITIONAL AGRI- ****** 4860 ****** CULTURAL POWER 4870 H = 0.0: N = 6 4880 Z8 = 0.0: FOR M8 = 1 TO N: GOSUB 4980 4890 28 = 28 + Q: M8: Z8 = Z8 / N 4900 H6 - H4 + Z8 4910 IF H6 >= 1.0 THEN 4950 4920 H8 = H8 + INT (H6 + 0.5) 4930 IF H8 >= H5 THEN 4950 4940 GOTO 4880 4950 RETURN 4960 END 5.2.1 Time, Cost and Seasonal Capacity Analysis ****** The time period of this model extends over six wet seasons from 1980 to 1985. This time span is chosen with regard to the estimate of tractor service life of six years (Directorate of Agricultural Mechani- zation, 1977-1978). The time analysis for the seasonal working hours and days computed from the equations: ED = (AD + 7) - (DL x RN) EH = (MH + c: x Q) where ED = potential suitable days in a season (in days) AD = allowable days based upon the irrigation schedule (in days) (5.4) (5.5) 74 DL = delay caused by late start of rain (in days) RN = number of heavy rainfall days (more than 40 millimeter/day) which might prevent tillage operations EH potential available hours in a season (hours) ME = mean Of working hours per day (hours) 0 = standard deviation for working hours per day (hours) Q = random probability number (see equation 5.1) 7 = number of extended days from the irrigation schedule commonly allocated in the study areas The potentially suitable days and available hours in a season are meant to represent the actual time spent in the rice field for rice land preparation, including idle time. The effective days and hours are the actual time spent for doing physical land preparation activi- ties, excluding idle time. The effective hours are used to predict the machine performance efficiency and capacity. The suitable days, and available hours are used for estimating cost of land preparation. Delay (DL) and rainfall (RN) are random variables computed from the subroutine Of random probability logic to obtain seasonal waiting time values. The capacity Of each technology in each season is pre- dicted by dividing the total effective working hours per season by field capacity (hours per hectare). Land preparation custom ratesiIIthe villages studied were paid for a day or bau contract basis. (Note: 1 bau = 0.7 hectare). Area based contracts were more commonly practiced for tractor operations, while 75 traditional land preparation was paid for on a day basis. The alge— braic expressions used for these two cost models were: MLCOST BLCOST TRCOST where MLCOST CAP HD 10,000 CM BLCOST TRCOST RATE (CD + CD x RATEl) + (CM + CM x RATEl) x 10,000 / CAP x HD (5.6) (CD + CD x RATE2) + (CM + CM x RATE2) x 10,000 / CAP x HD + RES x HLCOST (5.7) COSTBAU / 0.7 + (CM + CM x RATE3) + RES x HLCOST (5.8) manual labor cost (rupiah/hectare) capacity of technology (square meter/hour) random hours worked per day (hours) conversion number from square meter to hectare cost of meals for Operators (rupiah/day) cost of land preparation for bullocks (rupiah/ hectare) cost for land preparation for tractor (rupiah/ hectare) rate of annual increase of cost or price (Z/year) A11 variable values subject to change over time are up-dated each season using the discrete arithmetic method. 5.2.2 Farm Level Subsystem Model The logical path Of this subsystem followed the least cost preference and maximum delay constraint considerations. Rice farmers attempted to 76 select the least cost for land preparation in the sequential order of manual labor, bullocks and tractor power. Since the tractor tillage pricevuu;paid based upon unit areas, farmers gave attention to quality of the tillage operation with specific regard to depth of cut and the residual untilled land left on the plot corners and levee sides. The unit cost of land preparation of manual labor and bullocks, which were usually in terms Of rupiah per day,were estimated from farmers' responses about the past year's experiences. The area rate for these two techniques may vary from season to season depending on the degree of effectiveness during the working hours per day and the avail- able days per season; e.g., when during a full day of rain, farmers would only serve meals. This traditional contract does not apply to tractor Operators. Full payment is, however, made to manual labor or bullock teams whether they work a full day or not. Hours worked per day for these two techniques are a function of the number of working days per season. A farmer's decision as to the type of tillage operationdepended on the least cost and the availability of traditional technology. If, for instance, bullock cost for land preparationvnnslower, but the number Of waiting days exceeded the maximum tolerable days, the farmer looked for alternative technologies. The tolerable maximum waiting days varied depending on the customary practices in each village. Information regarding the demand for tractor land preparation can arrive relatively quickly at the tractor owners outside the village through the tractor owner association. Since tractors are mobile, it is assumed in this model that tractors have a greater degree of ability to serve 77 in land preparation than traditional methods. Random probability and waiting day subroutines are used to predict the possible waiting days for the availability of traditional power in this model. The simulation model outputs for farm level subsystems are in the form of: 1) labor used (man days/hectare), 2) labor and technological capacity (hours/ hectare and hectare/season), 3) the first and second pass of land pre- paration costs along with the total cost (rupiah/hectare), and 4) num- ber of waiting days and possible combination of technologies chosen which provide the total least cost information. 5.2.3 Tractor Owner Subsystem Model The mathematical model for analyzing the tractor owning costs and returns are presented by equations 5.9 to 5.12. The subsystem will predict the comparative profitability of tractor ownership under different purchasing arrangements and Operating conditions. The tractor returns from the tillage operation are based on land prepared. Therefore, tractor owners attempt to Operate their tractors as many hours per day as possible. The tractor expenses are predicted by Garmo and Canada, 1973; Kepner, g£_§l,, 1978. ANCOST = P x (i-e)(l+i-e)n/((l+i-e)n—l) (5.9) OPCOST = FCOST + OCOST + RMCOST + OPWAGE (5.10) TOTCOST = (ANCOST/X + OPCOST) x CAP (5.11) where i = annual interest rate (%) e = annual rate of inflation (%) ANCOST = annual (fixed) cost (RP/season) 78 P = purchasing price (RP) OPCOST operational cost per hour (RP/hour) FCOST = cost of fuel (RP/liter) OCOST = cost of oil (RP/liter) RMCOST = cost for repair and maintenance (RP/hour) OPWAGE = operator wage (RP/hour) X = number of effective working hours/wet season of 1980 CAP = capacity (hours per hectare) Diesel fuel and oil prices at the village level were RP 53 and RP 500 per liter respectively (December 1980). Repair and maintenance costs could not be Obtained accurately, because many tractors were new and tractor owners did not have sufficient information yet. Therefore, the level of repair and maintenance cost was estimated at 1.2 percent per 100 hours worked times 90 percent of the purchase price. Of the total of 1.2 percent, 0.8 percent was assumed for spare parts and 0.4 percent for mechanic's wages (Hunt 1973; Subdirectorate of Agricultural Mechanization, 1977-1978). The wet season fixed cost was estimated as being equal to the dry season fixed cost or in other words equal to 50 percent of the annual fixed cost. Annual fixed cost for capital recovery was computed from a uniform series payment. There was no tax, insurance or shelter costs for small tractors. The Operational (variable) cost consists of fuel, Oil, repair and maintenance costs and operator wages. The Operator wage is paid as a 79 percentage of gross return. Therefore, the operators are stimulated to work for more hours/day. The tractor net earnings in the wet season of 1980 was computed by subtracting total expenses from the customary rate paid by farmers. 5.3 ASystems Simulation Naylor (1960) defined simulation as the Operation of a model that represents an actual world system. Manipulation of the system inputs makes it possible to simulate the system's behavior under given assump— tions. Simulation models are best at providing an optimal range of in— formation rather than a single Optimal point. In the present modeling, each village was simulated individually along with the general or average condition. Since the input in each village varies, the generalized similation output should not be interpreted as an actual representation of the values of the variables being studied in each village. After the simulation model was verified, the sensitivity analysis was performed using average values of variables being studied. The generalized simulation was made to test the system modeling sensitivity under various levels of controllable input data. The degree of alterations of the system inputs were based upon the means or rates plus or minus the standard deviation or otherwise specified. The percentage and maximum probabilities of the delay of commence- ment of rain and waiting day estimations were Obtained from key farmers, sample farms and "PPL", the Field Extension Workers, in the study areas. This information was reconfirmed with the district irrigation service. In Jatiragas, for example, the probability of a delay of rainfall com- mencement is 75 percent and usually the maximum days of this delay does 80 not exceed 12 days. The probability Of obtaining manual labor eunibullocks in Tambah Dakan, for instance, is 66 percent and farmers, in this village, usually cannot wait for the availability of traditional power more than five days. Heavy rainfall which may prevent tillage Operation in Sukatani village, to give another example, is around 2.364 days, with standard deviation of 1.851 days. Specific exogenous data on manual labor migration from and to the villages were not available. Therefore, their numeric values are extrapolated from the population growth rate at the village level. The dynamic population of rural dwellers and bullocks for simulation inputs were derived from Table 4.1 of Chapter 4. Some environmental inputs related to financial analyses are presented in Table 5.2. Table 5.1 Annual rates of change for components Of the financial analyses. Item Rate of increase Manual labor wage rates1 11% Bullocks wage rates1 11% Tractor purchase price1 14% Retail price of rice (December)2 21% General index price (at rural area)1 15% General inflation (at national level)3 18% 1Central Bureau of Statistics (in Berustein,l980 ) 2Bustanil Arifin (National Logistics Board, 1978) 3Central Bureau of Statistics, 1978 81 The rate of increase of the variables presented in Table 5.2 were obtained by regression analyses using data presented in Appendix 2. Other system inputs, such as fuel and Oil consumption, hours worked per day, and land resources were derived through statistical inferences as discussed in Chapter 4. Sensitivity analyses were run three times to test the responsiveness of the simulation model to various input values derived from the average of village data. Critical controllable and parameter input components to be varied for these sensitivity analyses are 1) number of tractor and bullocks, 2) tractor prices, rate of interest, inflation and life expectancy Of tractor, and 3) hours and days worked available per season. The combina— tion and levels of numerical values of those critical components for the three sensitivity analyses are presented in Table 5.3. Table 5.2. Examples of input values for sensitivity analyses Item Sensitivity Analysis 1 2 3 Number Of tractors (units) 23 39 7 Number of bullocks 58 9 107 Tractor price (RP 1000) 2300 2300 2300 Rates of inflation (%/year) 15 15 18 Rates of interest (%/year) 10.5 10.5 12 Hours worked/day (hours) (a) manual labor 9.0 7.5 7.5 (b) bullocks 6.7 5.8 5.8 (c) tractors 13.3 11.6 11.6 Available days 35 30 30 Tractor life expectancy (years) 6 6 6 Note: Run 1, 2 and 3 indicate the run number under different input values 82 The numerical values of column 2 in Table 5.3 are obtained from the average data of eight villages discussed in Chapter 4, while figures in column 1 and 3 (run 1 and 3) are calculated based upon column 2:: the standard deviations. These variation levels of inputs are assumed close to the variation may exist in the real world. 5.3.1 Systems Simulation Inputs The uncontrollable exogenous inputs were simulated through random probability and waiting subroutines. The commencement and number of days of heavy rainfall were calculated statistically from the daily rainfall records over the past 31 years (1948-1980) at the Sukamandi rice seed farm which is located almost at the center of the villages studied. The values Of these variables are presented in Table 5.4. Table 5.3. Probability of delay, wait and heavy rainfall. Delay Wait Heavy Rainfall Village Prob. Max. Prob. Max. Means SD % days % days days days Jatiragas .75 12 .66 5 2.364 1.851 Jatisari .75 10 .75 7 2.364 1.851 Rawa Gempol .80 12 .58 4 2.364 1.851 Sukatani .70 15 .60 4 2.364 1.851 Bojong T. .80 14 .50 7 .535 .589 K. Anyar .70 14 .70 5 2.364 1.851 Mariuk .75 14 .50 4 .535 .589 T. Dahan .90 10 .66 5 .535 .589 Average: .77 13 .62 5 Standard Dev.: .06 2 .09 1 Source: Sukamandi Seed Farm (1980). 83 5.3.2 Systems Simulation Outputsland Discussion The systems simulation output of the farm level subsystems‘were 1) technical performance, which was measured in terms of technological capacity (hours/hectare; hectare/man day and hectares/season), 2) costs of the first and second tillage passes along with their total cost (Rupiah/ hour, Rupiah/hectare), 3) labor utilization, including manual labor for tillage operations on the corners and levee sides (man days/hectare), and 4) number of waiting days along with a combination of land prepara- tion techniques to produce possible least cost. The relationship between capital substitution for labor in land preparation is illustrated by isoquant curve. The simulation outputs of the tractor owner's financial subsystem were 1) expenses for variable (Operating), and seasonal (fixed) costs (Rupiah/hour, Rupiah/hectare), and 2) earning (Rupiah/hour, Rupiah/hec- tare) along with profitability levels (ratio of total present worth cost and earnings). Total and effective working hours for land preparation in a season WEIEIflmortant factors associated with the simulated technological capacity as well as financial analysis. The simulated outputs of effective hours per season are presented in Table 5.4. In this simulation model the available time due to irrigation schedule, delay of the commencement of rainfall and due to heavy rainfall along with periods necessary for softening rice land have been incorporated. The effective hours worked were simulated each day within the effective days of a season. The effective days were also obtained through the random probability subroutine. 84 Table 5.4. Simulated effective working hours in wet season, 1980 (hours) Village Manual labor Bullocks Tractors Jatiragas 204 149 497 Jatisari 195 162 373 Rawa Gempol 212 166 309 Sukatani 206 148 488 Bojong Tengah 223 208 336 Karang Anyar 250 194 343 Mariuk 246 172 418 Tambak Dahan 236 188 385 Average: 221 173 394 Standard deviation: 21 22 69 Coefficient variation: 9 13 18 Table 5.4 shows that tractor technology made possible to double the effective working hours in a season compared to bullocks technology and one and three fourths of the manual labor. The hourly manual labor and bullocks input per hectare in the eight villages showed a moderately high variation, with a coefficient of varia— tion of 21 percent and 26 percent respectively (see Table 5.5). The variation of time required for tractors to complete land preparation per hectare in the eight villages was the most stable with a coefficient Of variation equal to 13 percent. The hectarages capacity per season showed moderate variation for all these technologies with coefficients between 22 percent and 25 85 Table 5.5. Simulated capacity Of manual labor, bullocks and tractor technologies for wet season 1980 (two passes Of tillage Operations). Village Hours/hectare Hectares/season ML BL TR ML BL TR Jatiragas 368 61 13 0.55 2.4 38.3 Jatisari 319 47 13 0.60 3.4 27.6 Rawa Gempol 231 43 14 0.91 3.9 22.1 Sukatani 262 48 15 0.78 3.1 33.2 Bojong Tengah 257 43 17 0.87 4.8 19.5 Karang Anyar 220 51 13 1.13 3.8 29.5 Mariuk 383 62 15 0.64 2.8 27.9 Tambak Dahan 436 61 18 0.70 3.1 21.2 Average: .309 52 15 0.77 3.4 27.4 Standard Deviation: 63 8 2 0.19 0.7 6.4 Coefficient Variation: 21 26 13 25.0 22.0 23.0 percent. In general, seasonal capacity of tractors was almost 36 times greater than manual labor and eight times greater than bullocks technology. In summary, tractor tillage required 15 man hours per hectare for two passes (or about 105 to 127 horsepower hours per hectare) and the seasonal capacity was 27.4 hectares. While manual labor and bullocks technology requires 309 and 52 hours per hectare respectively and their seasonal capacities were 0.74 and 3.4 hectares respectively. These 86 results indicate that two wheeled tractors produce a large increase in (partial) labor productivity for rice land preparation. The simulated proportions of rice land prepared by manual labor, bullocks and tractors in each village are presented in Table 5.6. Table 5.6. Simulated proportion of rice land prepared by manual labor, bullocks and tractors, wet season, 1980.1 Village {:::1(;:§e Sioportional é?) prepared 3% Jatiragas 456 9 10 81 Jatisari 425 24 14 62 Rawa Gempol 850 6 6 88 Sukatani 683 16 10 74 Bojong Tengah 908 48 22 30 Karang Anyar 865 56 11 33 Mariuk 1111 9 27 64 Tambak Dahan 1138 61 21 18 Average: 804 29 15 56 Standard Deviation: 268 23 7 26 Coefficient Variation (%): 33 79 49 46 1Assumingnoimport and export of manuallabor,tnfllocks and tractors. These proportions indicate great variation with coefficients of varia- tions of all technologies above 30 percent. In Jatiragas, Rawa Gempol, Sukatani villages manual labor was already displaced by tractor technology (for detailed simulated results see Appendix 19). The proportion Of rice land prepared by manual labor 87 in these three villages was mostly the residual untilled land after bullocks and tractor Operations. In Jatisari, Karang Anyar, Mariuk and Tambah Dahan villages part of manual labor was displaced by tractors while in Bojong Tenhah there was a shortage of agricultural power. It was assumed that there was no migration of manual labor into this village. In this simulation 65 percent to 70 percent of all draft animals were assumed to be participating in land preparation. Further study in this area appears necessary, as one of the national development criteria is related to the increase of employment, while the present simulation indicates labor displacement in some of the villages studied. The amount of rice land prepared per man day is presented in Table 5.7. The variation of labor productivities Of all technology is moderately high with coefficients of variation 27 percent, 23 percent and 29 percent for manual labor, bullocks and tractor power respectively. The average labor capacity with tractor technology OfO.8493 hectare/ man day is 34.6 and 7.9 times greater than manual labor and bullocks respectively. The simulated labor utilization (man days per hectare) in the wet season of 1980 is presented in Table 5.8. Additional manual labor for tillage operations of the residual land was incorporated into the labor requirements for bullocks and tractors. In summary, based on simulated data for the wet season of 1980, the ratio of labor utilization for the three technologies were manual labor % bullocks % tractor equivalent to 9.6 % 2.6 % 1. This ratio implies that tractorized farms may save 9.6 times man days of manual labor or 2.6 times bullock man days for each tractor man day utilized. 88 Table 5.7. Simulated amount of rice land prepared per man day (two passes of tillages). Labor capacity Village (hectares/man day) Ratio ML BL TR BL/ML TR/ML TR/BL Jatiragas .0192 .0803 1.1511 4.2 59.9 14.3 Jatisari .0212 .1174 .8303 5.5 39.2 7.1 Rawa Gempol .0332 .1330 .7085 4.0 21.3 5.3 Sukatani .0312 .1057 1.2603 3.4 40.4 11.9 Bojong Tengah .0301 .1617 .5867 5.4 29.5 3.6 Karang Anyar .0392 .1272 .8912 3.2 22.7 7.0 Mariuk .0203 .0929 .7538 4.6 37.2 8.1 Tambak Dahan .0230 .0994 .6132 4.3 26.6 6.2 Average: .0272 .1147 .8493 4.3 34.6 7.9 Standard Dev.: .0073 .0259 .2438 .8 12.6 3.5 Coef. Variation (%): 27 23 29 19 36 44 Simulation outputs for cost of land preparation in the wet season of 1980 are presented in terms of Rupiah per hour and Rupiah per hectare in Table 5.9 and 5.10 respectively. In this simulated hourly cost of land preparation, additional manual labor in bullocks and tractor Operations, idle time except machine idle time and prices of meals, snacks and tobacco have been incorporated in the farmers' costs. Table 5.9 shows that the variation in the average farmer's cost for manual labor, tractor and tractor owner's costs had lower variation as indicated by coefficients of 10 89 Table 5.8. Simulated total labor utilization for two tillage Operations (man day/ha).1 Village Manual labor Bullocks Tractors Jatiragas 52.0 14.5 4.9 Jatisari 47.2 10.4 5.1 Rawa Gempol 30.9 8.6 3.4 Sukatani 32.0 10.8 2.9 Bojong Tengah 33.3 7.1 4.4 Karang Anyar 25.5 8.8 2.8 Mariuk 49.3 12.7 5.1 Tambak Dahan 43.5 11.9 4.3 Average: 39.2 10.6 4.1 Standard Deviation: 9.9 2.4 0.9 Coef. Variation (%) 25 23 23 lManual labor for tillage operation on residual untilled land was included. percent to 13 percent. Bullocks hourly cost varied somewhat higher with a 17 percent coefficient of variation. Hourly costs and levels of labor utilization will be used to construct isoquant curve (presented in Figure 5.5). The simulated land preparation costs per hectare for two passes are presented in Table 5.10. The farmers' costs included meals for the operators and costs of additional manual labor for bullocks and tractor Operations. 90 Table 5.9. Simulated hourly total costs of land preparation in the wet season Of 1980 (Rupiah/hour).l Farmers' costs for 2 Owners' costs Village Manual labor Bullocks Tractors Jatiragas 106 433 1693 844 Jatisari 113 364 1821 771 Rawa Gempol 95 307 1702 941 Sukatani 96 345 1584 666 Bojong Tengah 91 314 1325 917 Karang Anyar 81 255 1546 965 Mariuk 91 269 1448 782 Tambak Dahan 97 366 1342 1002 Average: 96 332 1558 861 Std. Deviation: 10 58 178 115 Coef. Variation(%): 10 17 ll 13 1Average of the first and second operations. 2Tractor owners' cash return is presented in Table 5.10. Farmers' costs of land preparation costs per hectare for manual and bullocks varied considerably between villages with coefficients of variation of 26 percent and 27 percent respectively. The cost for tractors paid by farmers was quite stable with a low (4 percent) coeffi- cient of variation. In most villages (except in Jatiragas and Tambah Dahan) the farmers' cost for bullock tillage was lower than for trac- tor. In Rawa Gempol, Sukatani and Karang Anyar, where manual labor costs were also lower than tractor costs. The lowest tractor owners' 91 Table 5.10. Simulated costs Of land preparation per hectare, two passes for wet season, 1980 (x RP 1000). I Village Manual laEOEmeBEllggizs Tractors Owners' COStS Jatiragas 45.4 39.7 31.0 11.0 Jatisari 42.7 26.1 31.6 10.2 Rawa Gempol 26.1 21.3 29.9 13.1 Sukatani 28.9 25.7 30.1 9.8 Bojong Tengah 28.3 20.1 28.8 15.5 Karang Anyar 21.8 19.9 27.8 12.9 Mariuk 42.0 25.3 31.0 11.7 Tambak Dahan 36.7 34.7 30.6 18.1 Average: 34.0 26.6 30.1 12.8 Std. Deviation: 8.8 7.1 1.3 2.8 Coef. Variation(%)26 27 4 22 lTractor owners' cash returns were RP 27,143 per hectare for each village except in Bojong Tengah and Karang Anyar which were RP 25,712 and RP 25,000 per hectare respectively. cost was found in Sukatani, possibly because this village had a longer experience with tractorization. Also operator wages were lower and longer hours were worked per day and per season (see discussion in Chapter 4, and Table 5.4). On the other hand, tractor owners' cost in Tambak Dahan was the highest among the villages. Less Operator experience, highest tractor purchase prices and lowest tractor utiliza- tion were found in Tambak Dahan. These factors apparently caused the tractor owning cost to be the highest in Tambak Dahan. Table 5.10 further indicates that in the wet season of 1980, tractor owners generally made great profits, their average return being almost two times their costs. 92 The simulated profitability index of the tractor owners in each village is presented in Table 5.11. This simulated result was obtained with the assumption of a 10.5 percent interest and 18 percent infla- tion rate, along with tractor life expectancy of 6 years. Table 5.11. Tractor owners' earnings, costs and profitability index, (earnings/costs), wet season 1980. Income from Total cost custom land for land Profitability Capital cost preparation preparation index per season per season per season (earnings/ Villages (Rp 1000) (Rp 1000) (Rp 1000) costs) Jatiragas 149.9 1,039.6 421.3 2.47 Jatisari 116.4 749.1 281.5 2.66 Rawa Gempol 149.8 599.9 289.5 2.07 Sukatani 125.3 901.1 325.4 2.77 Bojong Tengah 144.1 501.4 302.2 1.66 Karang Anyar 173.6 737.5 380.5 1.94 Mariuk 124.0 757.3 326.4 2.32 Tambak Dahan 188.8 575.4 383.7 1.50 1Figures are rounded. Further study to include dry season financial analysis is neces— sary in order to obtain an overall review of the profitability. A simulated result of locally manufactured tractors compared to imported tractor performances are presented in Table 5.12. 93 Table 5.12. Simulated locally manufactured and imported tractor performances for rice land preparation (two passes, wet season, 1980). Item Local tractor Imported tractorl (7~8.5 hp) (7~8.5 hp) Capacity Hours/hectare 16 15 Hectares/season 19.8 27.4 Hectares/man day .59 .85 Total labor input2 Man day/hectare 6.1 4.1 Operating costs Rupiah/hour3 (a) 623 847 (b) 842 876 Rupiah/hectare 11013 12800 Purchase price (RP 1000) 1650 2354 Returns/costs 2.465 2.125 Hours worked In a season (180) 324 394 In a day 9.7 11.6 1Average for eight villages. 2Manual labor for tillage Operation on residual untilled land is included. 3 tillage Operations. Figures are rounded to the one—tenth. 5Earning/cost. Total cost (fixed and variable costs); a, b are the first and second 94 Table 5.12 shows that the hourly field capacity and hectares/man day of locally manufactured tractors was 6.7 percent less than that of imported tractors. As has been discussed previously, this was due to technical problems with the steering mechanism. The capacity of imported tractors in Bojong Tengah and Tambah Dahan, however, was shown to be 6 percent and 12 percent lower than that of locally manufactured trac- tors. In the near future, the locally produced tractor will have a better acceptance for the following reasons: 1) a built-in steering clutch will improve maneuverability of the new model, 2) the simplicity of design was more appropriate for tractor Operators, few of whom have more than an elementary education, 3) plowing was more preferable for the first-over tillage Operation than rotary tillage, and 4) ease of repair (power transmitting devices, in particular) and better availability of spare parts. Hours worked per season is an important function of own- ing cost. Although the seasonal hours worked for the imported tractors were 22 percent higher, the return per hectare to the owner for local tractors was 10 percent less. This was due partly to a lower (30 per- cent) purchase price. A simulated situation with regard to the farmers' preference for a combination of land preparation techniques based upon least cost opportunity and maximum waiting time for the availability of traditional agricultural power is summarized in Table 5.13. In Jatisari, for in- stance, the projected land preparation cost for traditional power was RP 24,800/ha and the maximum number of days waited was four days. In Bojong Tengah, the lowest cost for the first tillage operation was with bullocks at a cost Of RP 12,400 per hectare, however, farmers would not 95 Table 5.13. Simulated necessary waiting days for traditional agri- cultural power and least cost Opportunity (wet season, 1980). Technology Total Cost Village ML BL TR wait (days) RP 1000 Jatiragas (a) (17.3) (b) (13.7) 0 30.1 Jatisari 2 (a) (14.2)(2) (b) (10.6>(2) 4 2““8 Rawa Gempol 3 (b) (11.0)(1) ' Sukatani (a) (11.8)(*) (16.3) 1 28.7 (b) (12.4)(1) Bojong Tengah (a) (12.4)(*) (16.4) 1 24.9 (b) ( 8.5)(1) Karang Anyar (a) (12.2)(*) (15.4) 0 27 7 (b) ( 7.5)(*) (12.3) ' Mariuk (a) (17.1)(*) (17.2) 0 31.0 (b) ( 8.2)(*) (13.8) Tambak Dahan (a) (20.4)(*) (16.8) 0 30 6 (b) (10.0)(*) (13.8) ' 1"a" and "b" indicate the first and second tillage operation respectively. 2 and waiting days respectively. 3 labor and bullocks exceeded the maximum waiting days. The figure in the first and second parenthesis are cost (RP 1000) The "*" indicates the waiting days for the availability of manual 96 wait the necessary seven days for the availability of bullock power. Therefore, the first tillage operation was with tractors. The combi- nation of tractors for the first pass and bullocks for the second pass would result in a cost Of RP 24900/hectare with one day of waiting for bullock power for the second tillage Operation. The sensitivity tests were made to identify the degree model response to various input variables. The examples of results are presented in Table 5.14. Table 5.14. The sensitivity analysis results. of this these test Sensitivity analysis Item 1 2 3 Capacity hours/hectare 15 15 15 hectare/season 29.5 26.6 26.6 hectares/man day 0.89 0.80 0.80 Total labor input for land preparation (man day) 3.6 4.5 4.3 Farmers' cost wet season 1980 (RP 1000/ha)l 30.3 30.6 30.6 wet season 1985 (RP 1000/ha) 60.5 60.1 69.2 Owners' cost wet season 1980 (RP 1000/ha)l 12.1 10.5 13.0 wet season 1985 (RP 1000/ha) 20.5 18.0 23.8 Note: Sensitivity 2 was used as basic run; in sensitivity 1, hours worked per day were altered; alteration of interest and inflation rates were made in sensitivity 3 (see Table 5.3). 1Projected cost. 97 Table 5.14 shows that tractor capacity (hectare/season or hectare/ man day) increased 11 percent when the hours worked per day increased with one unit of standard deviation value. Total labor input for land preparation (man day/hectare) decreases 20 percent. Therefore, this system modeling for tractor operation was not sensitive to hours worked per day, with reference to tractor capacity. However, labor input (man day/hectare) was sensitive to hours worked/day. Farmers' cosr (in the wet season of 1980) does not respond much to changes of interest and inflation rates. The projected cost in 1985 indicates sensitive response to changes of interest and inflation rates. Tractor owners' costs in the wet season of 1980 and 1985 increases 20 percent to 29 per- cent due to changes of interest and inflation rates. This means that the simulated tractor owners' cost model is sensitive to changes Of interest and inflation rates. CHAPTER 6 CONCLUSIONS From the data presented above, it can be concluded that: 1. Small two wheel tractors (7.0 to 8.5 horsepower) greatly increase labor capacity in rice land preparation in the Northern coastal plains of West Java. The tractor time required to complete two passes of tillage Operations per hectare (15 hours) was significantly less than the time required for tillage as performed by either bullocks or manual labor (52 and 309 hours per hectare respectively). 2. The simulated amount of land prepared for each technology per day in the wet season of 1980 was 34.6 and 7.9 times larger for tractor than manual labor and bullocks respectively. The hectarages Of land prepared for each technology per day were 0.0272, 0.1147 and 0.8493 hectare for manual labor, bullocks and tractor technology. 3. For both the first or the second pass of tillage Operations, tractors travelled two times faster than bullocks, with the average of both speed of 2.35 km/hr for the tractor and 1.17 km/hr for the bullocks. 4. The depth of the first manual labor tillage (6.2 cm) was significantly shallower than plowing with bullocks (10.7 cm) and tractor rototilling (10.3 cm). The difference Of tillage depth between bullocks and tractor was not significant. 5. The percentage of residual untilled land left by tractors was 7.3 percent of the total plot area and was significantly higher than 4.3 percent left by bullocks tillage. Within the plot size ranges of 98 99 0.09 to 0.21 hectare, the degree of tractor (machine) efficiency and the residual untilled land were weakly influenced by the plot dimension (the regression coefficient of determination was less than 0.1). 6. Number of effective hours worked per day was significantly greater for tractors than for either bullocks or manual labor; the averages being 11.6, 5.8 and 7.5 hours per day respectively. The simulated effective hours worked for the wet season of 1980 were 221 hours for manual labor, 173 for bullocks and 374 hours for tractors. Given these simulated hours worked per season, the capacity per season of manual labor was 0.77 hectares, 3.4 hectare for bullocks and 27.4 hectares for tractors. 7. The total labor utilized in land preparation including addi- tional manual labor for preparing residual land after bullocks and tractor operations was greatly reduced on farms using tractors as com- pared to farms using manual or bullock power. Tractorized farms utilized a total of 4.1 man days/ha for primary tillage as compared to 10.6 man days/ha on farms using bullocks and 39.2 man days/ha in farms using manual labor. 8. The total labor for land preparation plus related work derived from farm surveys was 25 percent, 30 percent, and 44 percent of the total labor input for rice preharvest activities for tractorized, bullocks and manual labor farms respectively. The total labor input for rice preharvest activities on farms using manual labor for land preparation (1979) was significantly greater than either bullocks or tractorized farms. There appeared to be no significant difference in preharvest labor input between bullock and tractor farms. The 100 average labor input for rice preharvest activities was 1274 hours for farms using manual labor, 779 hours for farms using bullocks and 741 hours for farms using tractors. 9. The simulated averages of land prepared at the village level (wet season, 1980) by tractors was 56 percent of the total rice land with the ranges of 18 percent in Tambak Dahan to 88 percent in Rawa Gempol village. The average proportion of rice land prepared by bullocks was 15 percent. A shortage of manual labor and bullock power for the wet season of 1980 at the village level was predicted in Bojong Tengah. There was some degree of simulated manual labor displaced by tractors in Jatisari, Karang Anyar, Mariuk and Tambak Dahan. Manual labor for land preparation in Jatiragas, Rawa Gempol and Sukatani villages had already been displaced by small tractors. In these villages manual laborers mostly prepared only the residual untilled land left by trac- tors and bullocks. 10. Simulated farmer costs of rice land preparation for the wet season of 1980 were generally lower for bullocks than either tractor or manual labor, except in Jatiragas and Tambak Dahan. Manual labor cost was the highest, with a total cost of RP 33,975 per hectare for the tillage operations. The farmers' costs for tractor and bullock land preparation (two passes) were RP 30,100 and RP 26,600 per hectare re— spectively. The tractor owners' costs for two passes of rototilling was RP 12,800 per hectare. The average simulated tractor owners' re- turns and costs ratio in the wet season of 1980 was calculated to be 2.12. 11. The capacity of locally manufactured tractors (7.0 to 8.5 hp, IRRI type design) was not significantly different from the imported 101 tractors. The field capacity was 16 hours per hectare to complete plowing and harrowing operations. The simulated labor utilization on farms using locally made tractors for land preparation was 6.1 man days. The returns and costs ratio was 2.46, which was higher than imported tractor profitability index. The price of locally manufactured tractors was RP 1,650,000 in 1980 (standard for rice field) which was 30 percent lower than the average price of imported tractors. 12. The level of rice production inputs (e.g., fertilizer, seed and chemicals) used were not significantly different between farms using manual labor, bullocks and tractors. Farms using tractors had signi- ficantly larger land holdings than those of bullocks and manual labor farms. 13. Farms using tractors and bullocks produced significantly higher rice yields per hectare as compared to farms using manual labor (survey results). The yield level at harvest moisture level in the wet season of 1979 was 4.1, 4.5 and 4.8 ton/hectare for manual, bullocks and tractorized farms respectively. The higher rice yield of 0.3 ton per hectare for tractor as compared to bullock farms was not statistically significant. In monetary terms, this 0.3 ton of rice difference was estimated at RP 21,000 (in 1979). CHAPTER 7 SUGGESTIONS FOR FURTHER STUDY This research compared alternatives of rice land preparation technologies during the wet season of 1980. It is suggested that dry season conditions be similarly studied to determine the complete financial analysis for tractor ownership on an annual basis. The farm survey indicated that there was no significant difference in total labor input for rice preharvest activities on farms using bullocks and tractors. It may be desirable to conduct research in this area by direct measurement at the farm level. The farm survey further indicated that farms using tractors produced higher rice yields per hectare, as compared to farms using traditional techniques. Further detailed study by direct measurement of this correlation between rice yield and land preparation technique at the farm level is suggested. 102 APPENDIX 103 uuuuuuuuuuuuuuuuuuuuu moo unusu::uununnunuununnunuuuuunnuunuu Hmuou ocmuo uuuuuuuuuu mma Inuulllln IIIIIIIIIIIIIII mHv Innuunnlunuuuuuu Hmuoe mm mm mm mm mm mp mm mm mm Hmuoulnsm m m m m m m m m a canon ansma m m m m m m m m m xsflumz m m m m m m m m m um>c¢ mcmumx m m m m m a m m m cmmcme mcoflom m o m m o m m a v flcmumxsm m m o m m m m m m HOQEwU m3mm m n m m m m m m m flummfiumh m s v m m m m o v mammuaumn m9 am A: m9 Am 42 m9 Am AZ mmmHHHu vcoomm momHHflu umuflm mommaaw> >m>u9m Eumm ucoemusmmmfi pamflm vocamuno mmHQEMm mo Hmnfisz " a xflocwma< 104 .NoH was songs .m mmeHa> as mxuoHHsn not uamuxm .NHH u mowm3 mxooaaon tam ponma vow: mo mums mcH NmH u wwmwuoafi wows“ Hmuwcow mo mum» mcfia .AOOOHV mm Ocm «N .mm maan .Cwmumcuom Eouw Omahamcmmm "mouaom OOH Om.m~m an Oq.OO ova ON.OOH oqa Om.mq «OH cam Omma NmH O0.00N OHH Om.mm mmH Om.me OMH O<.Oq mqa Nam mmma OOH Om.HON NOH Om.wq OOH OO.NwH OHH OO.mm OmH NON OnOH OOH O0.00H OOH O~.mq mHH OO.amH «Ha Om.mm HNH «mm mmma OOH O0.00H OOH ON.mq OOH Oo.mmH OOH O0.0N OOH OHN «50H I I I I I I I I I OOH Hmma axovcH Au:\mmv axmwaH Au:\amv mevcH Ap:\mmv wavcH Au£\amv ”nounwfimB mmm ham» mxooaaom uonma Hmscmz mxooHHom ponma Hmocmz xowcH Hmuocwo m mwmaaa> < mwmaaa> .Hm>oa mmeHH> wcu um cowumumaoum Ocma pow wwwms gonna mo omwouocfi mo oumm .N xfivcmoa< Appendix 3: 105 Ages of operators (years). No. Villages Technology Mean Range S.D. CV.(%) 1 Jatiragas Manual labor 37.5 29-45 6.8 18 Bullocks 36.0 27—50 8.8 24 Tractors 29 25-35 3.2 11 2 Jatisari Manual labor 37.9 20-60 10.7 28 Bullocks 34.4 25-40 5.0 15 Tractors 30.8 25-40 5.1 16 3 Rawa Gempol Manual labor 36.0 30-55 7.8 22 Bullocks 31.4 20-45 9.0 29 Tractors 30.1 24-42 6.6 22 4 Sukatani Manual labor 32.2 24-36 4.0 12 Bullocks 32.4 25-45 6.9 21 Tractors 28.1 23-35 4.7 17 5 B. Tengah Manual labor 39.4 25-55 11.2 28 Bullocks 34.8 25-50 8.4 57 Tractors 30.3 20-40 7.9 26 6 K. Anyar Manual labor 34.7 25-50 9.9 29 Bullocks 35.5 20-45 9.2 26 Tractors 28.9 25—37 4.3 15 7 Mariuk Manual labor 45.1 35-65 11.0 24 Bullocks 42.8 35-55 8.2 19 Tractors 30.3 25-40 5.1 17 8 T. Dahan Manual labor 47.6 30-60 11.0 23 Bullocks 43.3 30-52 7.7 18 Tractors 26.2 20-30 3.5 13 Note: SD = standard deviation; CV = coefficient of variation 106 Appendix 4: Operator experience (years). No. Villages Technology Mean Range S.D. C.V.(Z) l Jatiragas Manual labor 16.5 1-30 12.8 78 Bullocks 15.5 2-30 11.2 72 Tractors 2 0 1- 4 1.2 60 2 Jatisari Manual labor 17.3 1-40 12.0 69 Bullocks 14.2 1-20 7.4 52 Tractors 1.6 1- 3 0.7 44 3 Rawa Gempol Manual labor 20.0 10-40 8.8 44 Bullocks 13.9 5-25 7.3 52 Tractors 1.2 0.1- 2 0.9 75 4 Sukatani Manual labor 15.1 -25 8.0 53 Bullocks 14.4 3-30 9.9 69 Tractors 2 4 0.1- 5 1.5 61 5 B. Tengah Manual labor 23.8 10-40 10.6 45 Bullocks 14.8 9-30 8.4 57 Tractors 1.1 0.1-2 0.6 55 6 K. Anyar Manual labor 17.8 2-30 9.4 53 Bullocks 19.5 9-30 8.0 41 Tractors 1.8 l- 4 1.1 61 7 Mariuk Manual labor 24.3 2-50 13.3 55 Bullocks 25.6 15-40 9.8 38 Tractors 1.7 l- 3 0.8 47 8 T. Dahan Manual labor 27.0 10-40 14.2 53 Bullocks 24.3 10-35 9.3 38 Tractors 0.9 0.1- 2 0.6 67 Note: S.D. = standard deviation; C.V. coefficient of variation. 107 Appendix 5: Effective worked hours per day for the first Operation (hours). No. Villages Technology Mean Range S.D. C.V.(Z) 1 Jatiragas Manual labor 7.8 6.0- 8.5 1.2 15 Bullocks 5.2 4.3- 7.2 0.9 17 Tractors 15.4 9.5-24.0 4.9 32 2 Jatisari Manual labor 6.8 5.0— 9.0 l 5 22 Bullocks 5.2 4.5- 6. 0.8 15 Tractors 13.3 7.0-20.0 4.5 34 3 Rawa Gempol Manual labor 7.7 4 8- 9.5 1.6 21 Bullocks 6.7 4.0- 9.0 1.6 24 Tractors 10.0 4.0-12.5 3.6 36 4 Sukatani Manual labor 6.6 5.0- 8.0 1.1 17 Bullocks 5.1 3.8- 6.2 0.8 16 Tractors 9.6 7.0-16.5 3.0 31 5 B. Tengah Manual labor 7.4 4.7— 9.3 1.8 24 Bullocks 6.9 5.2- 8.0 1.2 17 Tractors 9.6 7.0-16.5 3.0 31 6 K. Anyar Manual labor 8.3 5.0-10.0 1.6 19 Bullocks 6.1 4.2- 7.5 1.2 3 Tractors 11.7 8.0-16.0 2.8 10 7 Mariuk Manual labor 8.0 5.5- 9.8 1 5 19 Bullocks 5.1 3. - 9.0 2.3 45 Tractors 12.8 9 5-19.0 3.6 28 8 T. Dahan Manual labor 7.9 5.5- 9.6 1.4 18 Bullocks 7.0 4.2-10.0 2.0 29 Tractors 10.1 7.0-15.5 2.6 26 Note: S.D. = standard deviation; C.V. = coefficient of variation. 108 Appendixe : Effective worked hours per day for the second operation (hours). No. Villages Technology Mean Range S.D. C.V.(Z) 1 Jatiragas Manual labor 6.3 4.2-10 6 2.0 32 Bullocks 4.7 3.5- 6 6 1.0 21 Tractors 12.8 11.3-15 0 1.3 10 2 Jatisari Manual labor 6.8 4.0- 9.0 1.6 23 Bullocks 5.7 4.5- 7.0 0.7 12 Tractors 9.3 6.0-11.6 2.2 24 3 Rawa Gempol Manual labor 7.7 5.0-10.0 1.6 21 Bullocks 4.9 4.2- 7.9 0.9 18 Tractors '9.6 6.2-13.2 2.5 26 4 Sukatani Manual labor 7.7 4.2- 9.5 1.7 22 Bullocks n.a. n.a. n.a n.a. Tractors 15.0 8.0-22.5 4.6 31 5 B. Tengah Manual labor 8.2 7.1- 9.2 0.9 11 Bullocks 7.1 5.6- 9 2 1.4 20 Tractors 10.3 8.0-12 5 0.7 7 6 K. Anyar Manual labor 8.9 8.0-10.0 0.7 8 Bullocks 6.9 5.5- 9.2 1.6 23 Tractors 11.4 8. -13.5 2.1 18 7 Mariuk Manual labor 7.7 6.0- 9 2 1.0 13 Bullocks 5.6 3.2- 9 5 2.3 41 Tractors 10.0 8.0-20 0 4.8 48 8 T. Dahan Manual labor 7.6 5.0- 8.0 1.1 14 Bullocks 5.2 4.5- 6.0 0.5 10 Tractors 11.4 10.5-16.5 2.9 25 Note: S.D. = standard deviation; C.V. = coefficient of variation; n.a. = not available. Appendix 7 = 109 Idle time for the first tillage operation (hours) No. Villages Technology Mean S.D. C.V.(Z) l Jatiragas Manual labor 1.7 0.2 14 Bullocks 0.5 0.3 58 Tractors 1.8 0.8 44 2 Jatisari Manual labor 1.2 0.9 75 Bullocks 0.4 0.1 25 Tractors 1.8 0.8 44 3 Rawa Gempol Manual labor 1.3 0.5 38 Bullocks 1.0 0.9 89 Tractors 1.6 0.9 60 4 Sukatani Manual labor 1.2 0.6 50 Bullocks 0.4 0.1 25 Tractors 1.4 0.8 57 5 B. Tengah Manual labor 1.5 0.8 53 Bullocks 0.9 0.4 44 Tractors 1.1 0.7 64 6 K. Anyar Manual labor 1.5 0.6 40 Bullocks 0.8 0.4 51 Tractors 1.9 0.7 37 7 Mariuk Manual labor 1.5 0.6 40 Bullocks 1.0 0.7 72 Tractors 2.5 1.2 49 8 T. Dahan Manual labor 1.8 0.8 44 Bullocks 1.4 1.2 86 Tractors 1.5 0.6 40 Note: S.D. = standard deviation; C.V. = coefficient of variation. 110 Appendix 8 : Idle time for the second tillage operation (hours). No. Villages Technology Mean Range S.D. C.V.(Z) l Jatiragas Manual labor 0.9 0.2-2.5 0.8 89 Bullocks 0.3 0.2-0.5 0.1 27 Tractors 2.3 1.2-3.3 0.7 30 2 Jatisari Manual labor 1.3 0 5-1.8 0.4 31 Bullocks 0.5 0 3-0.6 0.1 28 Tractors 0.5 0 3-0.6 0.1 25 3 Rawa Gempol Manual labor 1.5 0.2-2.0 0.6 40 Bullocks 0.4 0.2-0.6 0.1 31 Tractors 1.4 0. -2.5 0.8 59 4 Sukatani Manual labor 1.3 0.3-0.2 0.5 38 Bullocks n.a. n a. n.a. n.a. Tractors 2.2 1.0-3.5 1.0 45 5 B. Tengah Manual labor 1.7 1.0-2.5 0.5 29 Bullocks 1.6 0.6-2.5 0.8 48 Tractors 1.6 0.6-2.6 0.7 43 6 K. Anyar Manual labor 1.9 1.2-2 2 0.3 15 Bullocks l 4 0.7-3 0 0.8 58 Tractors 2.1 1.6-2 6 0.4 17 7 Mariuk Manual labor 1.8 1.2-2.5 0.4 22 Bullocks 1.0 0.2-2.5 0.8 85 Tractors 2.6 0 6-3 5 0.9 35 8 T. Dahan Manual labor 0.4 0.2-0.5 0.1 27 Bullocks 1.7 0.2—2.2 0.6 35 Tractors 1.9 0.2-3.3 0.9 47 Note: S.D. standard deviation; C.V. = coefficient of variation; not available. 111 Appendix 9 : (centimeters). Depth of tillage of the first tillage operation No. Villages Technology Mean Range S.D. C.V.(%) 1 Jatiragas Manual labor 6.08 5.91- 6.27 0.17 2.90 Bullocks 12.48 10.85-16.28 1.65 13.21 Tractors 10.71 9.00-14.l4 1.85 17.26 2 Jatisari Manual labor 6.41 5.72- 7.26 0.65 10.23 Bullocks 11.30 9.26-18.00 2.64 23.41 Tractors 9.95 9.53-10.26 0.30 3.06 3 Rawa Gempol Manual labor 6.41 5.81- 6.83 0.38 5.94 Bullocks 10.79 9.00-11.91 0.92 8.53 Tractors 10.25 8.34-11.46 1.16 11.36 4 Sukatani Manual labor 5.61 5.27- 6.25 0.37 6.66 Bullocks 11.23 9.56-12.00 0.86 7.66 Tractors 8.80 5.97-10.91 1.60 18.25 5 B. Tengah Manual labor 6.10 5.85- 6.31 0.20 3.26 Bullocks 10.95 8.26-12.37 1.27 11.61 Tractors 10.50 8.12-1l.92 1.20 11.48 6 K. Anyar Manual labor 6.16 5.48- 6.56 0.37 6.15 Bullocks 10.13 7.07-1l.92 1.53 15.14 Tractors 11.29 6.57-17.l4 3.81 33.73 7 Mariuk Manual labor 7.45 5.53- 9.40 1.25 16.88 Bullocks 9.79 7.80-11.26 1.16 11.91 Tractors 11.23 10.20-12.46 0.94 8.41 8 T. Dahan Manual labor 5.30 5.35- 5.66 0.35 6.71 Bullocks 9.08 7.93- 9.90 0.90 9.90 Tractors 10.06 9.13-11.26 0.89 8.87 Note: S.D. = standard deviation; C.V. = coefficient of variation. 112 Appendixlo ; Traveling speed of the first pass (kilometer/hr) No. Villages Technology Mean Range S.D. C.V.(Z) 1 Jatiragas Bullocks 0.48 0.34-0.55 0.06 12.43 Tractors 1.06 1.01-1.09 0.02 2.43 2 Jatisari Bullocks 0.49 0.35-0.78 0.14 29.33 Tractors 1.18 1.06-1.31 0.12 10.83 3 Rawa Gempol Bullocks 0.43 0.36-0.46 0.03 6.83 Tractors 0.83 0.78-0.88 0.04 5.25 4 Sukatani Bullocks 0.45 0.33-0.54 0.06 14.22 Tractors 0.73 0.39-1.39 0.29 39.43 5 B. Tengah Bullocks 0.47 0.36-0.53 0.06 13.51 Tractors 0.65 0.34-1.14 0.25 38.92 6 K. Anyar Bullocks 0.41 0.38-0.45 0.02 6.02 Tractors 0.83 0.65-1.06 0.11 13.53 7 Mariuk Bullocks 0.41 0.31-0.52 0.06 16.62 ' Tractors 0.81 0.50-1.25 0.34 42.05 8 T. Dahan Bullocks 0.39 0.32-0.44 0.05 14.14 Tractors 0.74 0.51-0.96 0.17 22.99 Note: S.D. = standard deviation; C.V. = coefficient of variation. 113 Appendix 1;; Traveling speed of the second pass (kilometer/hr) No. Villages Technology Mean Range S.D. C.V.(Z) l Jatiragas Bullocks 0.44 0.32-0.60 0.10 22.79 Tractors 0.73 0.57-0.86 0.10 14.92 2 Jatisari Bullocks 0.38 0.25-0.60 0.09 25.31 Tractors 0.95 0.48-1.22 0.23 24.88 3 Rawa Gempol Bullocks 0.40 0.36-0.45 0.03 8.39 Tractors 0.87 0.82-0.99 0.06 6.92 4 Sukatani Bullocks 0.37 0.28-0.44 0.06 16.95 Tractors 0.74 0.69-0.81 0.03 4.93 5 B. Tengah Bullocks 0.38 0.20-0.52 0.10 28.29 Tractors 0.88 0.71-1.13 0.16 19.01 6 K. Anyar Bullocks 0.48 0.37—0.64 0.08 18.29 Tractors 1.09 0.80—1.41 0.22 20.49 7 Mariuk Bullocks 0.38 0.27-0.48 0.07 19.27 Tractors 0.77 0.60-0.91 0.11 14.76 8 T. Dahan Bullocks 0.48 0.36-0.77 0.16 33.94 Tractors 0.69 0.60-0.79 0.07 10.30 Note: S.D. = standard deviation; C.V. = coefficient of variation. 114 Appendix 12 .Fhel and oil consumption (liter/hr). No. Villages Fuel/Oil Mean Range S.D. C.V.(Z) 1 Jatiragas Fuel 1 1.249 0.824-1.841 0.349 27 Fuel 2 1.183 1.020-1.428 0.156 13 Oil 0.047 0.031-0.068 0.009 21 2 Jatisari Fuel 1 1.486 1.120-1.761 0.272 18 Fuel 2 1.330 l.082-l.485 0.117 9 Oil 0.036 0.025-0.050 0.009 25 3 Rawa Gempol Fuel 1 1.192 0.964-l.471 0.189 16 Fuel 2 1.122 1.003-1.237 0.075 7 Oil 0.036 0.025-0.060 0.116 32 4 Sukatani Fuel 1 1.276 0.658-l.829 0.376 29 Fuel 2 1.174 0.870-1.409 0.183 16 Oil 0.049 0.024-0.081 0.024 48 5 B. Tengah Fuel 1 1.030 0.668-1.432 0.225 22 Fuel 2 1.067 0.705-1.500 0.374 35 Oil 0.030 0.027-0.032 0.004 14 6 K. Anyar Fuel 1 1.308 0.806-1.636 0.311 24 Fuel 2 1.109 0.631-1.562 0.312 28 Oil 0.040 0.026-0.350 0.010 25 7 Mariuk Fuel 1 1.199 0.909-1.500 0.237 20 Fuel 2 1.285 0.919-1.875 0.277 22 Oil 0.050 0.020-0.088 0.026 52 8 T. Dahan Fuel 1 0.968 0.698-1.286 0.199 21 Fuel 2 1.171 0.932-1.338 0.155 13 Oil 0.050 0.030-0.060 0.020 40 Note: S.D. = standard deviation; C.V. coefficient of variation. 115 Appendix 13 : Percentage residual untilled land after the first bullock and tractor operation (%).* No. Villages Technology Mean Range S.D. C.V.(%) 1 Jatiragas Bullocks 7.61 2.78-17.87 4.77 62.65 Tractors 7.61 2.63-13.19 3.76 49.41 2 Jatisari Bullocks 4.09 2.03- 6.48 1.59 38.99 Tractors 8.38 5.64-12.00 2.44 29.17 3 Rawa Gempol Bullocks 3.43 0.98- 6.94 2.02 58.98 Tractors 6.43 4.67- 7.90 1.12 17.48 4 Sukatani Bullocks 4.59 0.95- 7.40 1.88 40.94 Tractors 6.89 3.50-11.98 2.64 38.30 5 B. Tengah Bullocks 2.86 1.54- 5.20 1.20 42.07 Tractors 8.10 5.36-13.63 2.51 30.95 6 K. Anyar Bullocks 3.81 1.86- 6.22 1.70 44.54 Tractors 6.78 4.39-15.49 3.44 50.75 7 Mariuk Bullocks 3.96 0.64-12.11 3.42 86.57 Tractors 7.78 3.21-12.44 2.69 34.59 8 T. Dahan Bullocks 4.38 1.42-14.68 4.28 97.69 Tractors 6.11 3.06-10.50 2.72 44.56 Note: S.D. = standard deviation; C.V. = coefficient of variation; * = percentage of the rice field (jolat) area. 116 Appendix 14 : Estimated prices of hoe, bullocks and tractor (Rupiah, 1000, 1980). No. Villages Technology Mean Range S.D. C.V.(Z) 1 Jatiragas Hoe 1.592 1.000-2.500 343 21 Bullocks 300 200- 400 56 19 Tractors 2,380 l,950-2,700 230 10 2 Jatisari Hoe 1.481 1.000-2.000 330 22 Bullocks 270 210- 330 35 13 Tractors 1,850 1,600—2,450 790 43 3 Rawa Gempol Hoe 1.469 1.000-2.000 386 26 Bullocks 300 270- 350 19 6 Tractors 2,380 1,900-2,700 290 12 4 Sukatani Hoe 1.386 500-1.600 321 23 Bullocks 280 100- 400 130 46 Tractors 2,000 1,300-2,500 450 22 5 B. Tengah Hoe 1.744 1.500-2.500 382 22 Bullocks 270 200- 320 36 13 Tractors 2,290 1,300-2,800 600 26 6 K. Anyar Hoe 1.744 1.300-2.000 271 15 Bullocks 300 250- 320 24 8 Tractors 2,760 1,950-3,700 560 20 7 Mariuk Hoe 1.400 1.000-2.000 283 20 Bullocks 260 100- 350 65 25 Tractors 1,970 1,050-2,890 690 35 8 T. Dahan Hoe 1.473 1.000-2.500 461 31 Bullocks 310 300- 320 16 5 Tractors 3,200 2,180-3,600 543 17 Note: S.D. = standard deviation; C.V. coefficient of variation. Appendix 15 117 Operator wages (Rufriah/day for manual labor and bullocks; percentage (2) for tractor operators). No. Villages Technology Mean Range S.D. V-(%) 1 Jatiragas Manual labor 538.5 500- 600 50.5 9 Bullocks 1,656.2 1,000-2,000 301.0 18 Tractors 15.3 15- 16 0.3 2 2 Jatisari Manual labor 558.3 500- 600 51.5 9 Bullocks 1,533.3 1,500-1,600 48.5 3 Tractors 15.0 15-15 0.0 0 3 Rawa Gempol Manual labor 513.3 500- 600 35.2 7 Bullocks 1,366.7 500-2,500 431.6 32 Tractors 15.0 15-15 0.0 0 4 Sukatani Manual labor 542.9 500- 600 51.3 9 Bullocks 1,280.0 1,000-1,500 258.8 20 Tractors 11.6 10- 15 2.4 21 5 B. Tengah Manual labor 508.8 500- 650 36.4 7 Bullocks 1,933.3 1,000-2,600 627.8 32 Tractors 16.3 10- 25 4.0 24 6 K. Anyar Manual labor 494.1 400- 500 24.2 5 Bullocks 1,256.5 500-1,900 490.3 39 Tractors 14.3 12- 15 1.3 9 7 Mariuk Manual labor 505.9 500- 600 24.2 5 Bullocks 1,142.8 500-2,000 602.2 53 Tractors 16.7 10- 20 3.7 22 8 T. Dahan Manual labor 500.0 500- 500 0.0 0 Bullocks 2,142.8 l,500-2,000 363.1 17 Tractors 10.9 10- 12 1.0 9 Note: S.D. = standard deviation; C.V. coefficient of variation. Appendix 16. Simulated farmers' costs of land preparation per 118 hectare for the first and second operations, during the wet season of 1980 (Rp 1000). Manual labor Bullocks Tractors Village A1 Bl A1 31 A1 31 Jatiragas 27.7 17.7 19.2 20.4 17.9 13.8 Jatisari 32.1 10.6 14.2 11.9 17.9 13.8 Rawa Gempol 15.0 11.0 8.9 12.3 16.1 13.8 Sukatani 16.5 12.4 11.8 13.9 16.3 13.8 Bojong Tengah 15.5 12.8 12.4 8.5 16.4 12.3 Karang Anyar 14.3 7.5 12.2 7.7 15.4 12.3 Mariuk 26.1 16.0 17.1 8.2 17.2 13.8 Tambak Dahan 26.7 10.0 20.4 14.3 16.8 13.8 Average: 21.7 12.2 14.5 12.1 16.7 13.4 Std. Deviation: 7.1 3.3 4.0 4.2 0.9 0.7 Coef. Variation (Z): 33 27 26 35 5 5 1"A" and "B" are the first and second operations respectively. 2Figures are rounded. 119 Appendix 17. Simulated costs of owning tractor per hour (wet season, 1980).1 First operation Second operation Village Fixed cost (Rp/hr) (Rp/hr) (RP 1000/season) Variable Total Variable Total Jatiragas 149.9 613 870 560 818 Jatisari 116.4 508 712 626 830 Sukatani 125.3 479 673 465 660 Bojong Tengah 144.1 493 879 570 956 Karang Anyar 173.6 591 967 588 964 Mariuk 124.0 519 775 533 789 Tambak Dahan 188.8 525 984 561 1021 Average: 146.5 535 847 564 876 Std. Deviation: 25.1 47 115 49 121 Coef. Variation(%): 17 9 14 9 14 1Rates of interest 10.5%/year; inflation 18%/year; expected tractor service life = 6 years. 120 .mu.O he wowwaowuHse zn musoc :mE ou Owupw>coo .cwan >3 ocoOA «4.555 mw.mmo so.aeaa "fiance «.54 wq.ca ow.mm 5.00 ow.55 o~.qm m.ma mm.5 4H.Hq msomcmaamumaz 5.m oq.wN m~.oo II II m~.qa II II oo.oa mumu coaumwauuH «.mo oq.qm ~m.cm «.mq 50.x oo.o~ 5m.5m m~.Hq No.55 wcw5muam «.mm ma.co o~.cwa 5.54 Hm.¢w mw.a5H m.mq ma.qaa mq.em~ lwcaemmz c.5m w~.c 5o.~H m.oo 5o.qa q~.o5 H.eq qq.ma «H.m~ wcwmmmuu ace 0.5m w~.c 5o.~H m.mo 5~.ma wa.ma 5.0a 5a.m mq.ma meanaaauumu gamma o.m~ os.o~ mo.m5 c.0o am.qo M5.oa n.5m mq.5m om.ooH lwcauamaamcmue o.wm oo.wH mo.oq «.mm oH.5m mo.qm o.m~ mo.q~ mm.mm ate; eamau Hmcoauaeuc m.5q Hm.mw 5o.w5H w.qm mm.am 5m.wqa m.o5 m5.qm 50.55H wcamom 5.moH sq.oq o.mm w.wq 5H.oN mm.aq H.HN 5m.qw o.ooq coaumumamua cams o.~m Hq.qa H~.5~ 5.wm mo.¢a mm.5~ o.m mm.H 5m.mm summusz HmU mm cmmz ANV>U am cam: ANO>U am cmmz mfiuom umw>um£mua moan pom usacw uopmq OH xfivcmaa< 121 .m5.O >3 wcwxaawuase >3 muooz :45 04 Omuum>coo .cmEoB >4 mcoa4 44.444 44.544 44.4454 444444 4.55 45.44 44.55 4.55 45.4 54.45 5.45 44.4 44.44 4444444444442 4.45 54.45 45.55 4.4 45.5 44.45 II II 45.54 4444 4444444444 5.44 54.45 44.54 4.45 44.54 54.55 4.44 44.44 44.55 44454444 4.55 45.44 55.545 4.45 54.45 45.445 5.55 55.55 44.445 44444442 5.44 55.54 44.45 4.55 45.4 45.44 5.45 44.44 44.55 44444444 444 5.54 54.44 45.55 5.45 54.4 54.45 5.4 54.4 54.44 44444444444 44444 5.45 44.45 45.444 4.44 45.44 44.444 4.45 55.45 44.44 44444444444444 5.45 54.44 44.554 4.55 44.45 44.554 4.44 55.55 44.44 4443 44444 4444444444 II II 44.44 II II 44.445 4.45 54.444 44.454 444442 4.44 45.4 44.55 5.45 44.44 44.54 4.44 44.545 44.444 44444444444 4444 4.45 54.44 44.44 5.44 44.4 54.45 4.44 45.45 44.44 5444442 42444244 54.554 45.454 55.5554 444444 4.45 55.4 44.45 4 4 45.54 4.55 55.45 45.55 4444444444442 4.54 45.55 54.55 II II II II II 44.55 4444 4444444444 5.44 45.44 55.45 5.54 44.4 44.54 4.45 44.55 45.54 44454444 4.44 54.55 55.454 4.54 54.45 45.545 4.4 55.54 44.545 44444443 5.45 45.5 55.54 4.45 45.5 55.45 4.45 54.4 55.54 44444444 444 5.45 44.5 55.54 4.44 54.4 54.5 4.454 54.45 44.54 44444444444 44444 5.44 44.55 45.454 4.44 45.44 44.544 4.44 44.44 44.454 44444444444444 5.45 45.44 55.554 5.45 45.55 45.44 5.45 45.44 44.554 4443 44444 4444444444 II II II 5.44 54.44 44.55 5.45 44.45 45.544 444442 5.45 54.4 45.54 4 4 45.45 5.45 54.454 44.445 44444444444 4444 4.54 44.4 54.54 5.44 54.45 45.44 5.54 44.44 44.45 5444442 454>4 44 4442 454>4 44 4442 454>4 44 4442 444244 4344 HOuumHH xUOHHDm HO£NH ~4th Av.ucoov OH x4vcmma< 122 .05.0 50 00000000008 50 muaos awe Ou Omuum>co0 .cmso3 50 0coa4 50.000 00.005 40.0400 "0400.4. 0.00 O0.00 00.04 0.00 00.00 50.00 0.45 04.44 O0.00 wsomcm00mom0z II II O0.00 II II II II II II 0440 :000400440 0.54 00.00 40.500 0.00 O0.40 40.000 0.00 00.04 05.00 0c05muam 0.00 0O.44 00.050 0.00 40.00 55.050 0.00 00.04 05.000 40000003 0.00 04.0 05.00 0.000 00.00 40.00 0.00 04.00 00.40 0:000040 009 0.05 O0.40 O0.00 4.04 O0.0 00.00 0.00 00.00 50.04 0:050000400 00000 4.04 04.44 00.50 0.00 00.00 50.00 5.00 00.00 40.00 40:0ucm0amcmpe 0.00 00.44 05.000 0.50 00.00 00.00 0.000 50.000 O0.000 x4o3 00000 0m:O00000< II II O0.04 0.00 00.000 00.000 0.04 00.00 00.000 0:000: 0.00 40.0 00.40 0.00 00.00 00.04 0.50 40.00 00.000 cOHumumamua 0040 0.04 00.50 00.00 0.0q 00.00 04.44 0.00 00.00 05.04 5404452 m4 44 4442 454>0 45 4442 454>0 44 4442 m400 .44503 00 44004 4.444 44.545 44.4544 “44444 4.444 44.44 40.44 4.44 40.54 44.44 0.44 04.40 44.04 444444444444: 0.444 44.04 44.44 II II 44.44 II II 44.44 4444 4444444444 5.04 04.44 44.45 4.44 44.54 44.40 0.44 45.44 04.40 44444444 4.45 44.444 04.454 0.44 04.45 04.044 0.04 40.444 44.440 44444443 4.44 45.04 44.04 4.04 44.4 44.54 4.45 44.44 44.40 44444444 444 5.55 04.44 44.54 4.04 44.4 44.54 4.45 44.44 44.40 44444444444 44444 4.50 44.44 44.004 4.44 40.45 55.444 4.40 45.44 04.444 44444444444444 4.44 44.44 40.444 4.45 44.44 44.44 4.44 44.44 44.54 4443 44444 4444444444 0.00 44.044 00.404 0.40 00.004 45.504 0.04 05.54 44.00 044440 4.44 44.44 44.44 4.44 44.44 44.44 0.44 4.444 44.454 44444444444 4444 4.04 45.44 00.50 0.54 50.44 44.44 0.40 40.40 05.44 0444442 24444 4442<4 44.445 40.400 45.4444 "44444 0.05 54.44 44.44 0.40 40.4 54.44 4.44 45.4 44.04 444444444444: 0.55 40.04 44.44 4.044 44.05 45.40 II II 44.40 4440 4444404444 4.44 00.44 44.05 0.44 05.54 00.44 4.04 40.44 00.45 44404444 5.44 45.04 40.054 4.40 00.44 44.444 4.44 44.40 45.440 44444443 4.04 44.4 40.54 4.04 44.4 04.04 4.40 45.4 54.40 44444444 444 4.04 40.4 04.44 0.44 44.4 45.44 0.04 40.4 00.44 04444444440 04440 4.40 40.40 44.00 0.40 44.04 44.444 4.44 45.54 40.004 H4444444444444. 5.44 44.44 54.444 4.40 04.04 04.444 4.44 44.44 44.54 4443 44444 4444444444 II II 44.44 4.44 44.44 40.444 4.44 45.04 44.044 444444 0.40 40.4 04.40 0.44 44.04 44.544 4.44 40.04 44.044 44444444444 4444 4.40 44.44 44.54 4.44 44.00 44.04 4.04 50.44 44.04 0444442 400>0 44 4442 40040 44 4442 400>0 44 444: 4444 comm ca mmeumw zmx van mpmxuoz coamawuxm wamfim .Ammv vcmaas How wcfivsaoafi .mpmuom: uma mzmv awe ma mwmum>m cmE:mm<: .wcma mUflu HMuou msu Scum vama Hmsvwmmu cam uouumuu vcm xuoaaan An wmumamua vama wswuuwuumnsm >3 vwcflmunom .m:oaumumao acuumuu vam xuoaasn umumm vcma wmaafiucs mo mwMucmuuma com: vmmmn vmumefiummm .pm3oa Hmuauasowuwm Am>fiuu=vouav waamafim>m «0 “mafia: x muaumamu Hmcowmmw u con: cummn vmumanoamofi 124 mam e~m me hema I mom mafia Now Hmm uaNEouuxflaasw mama «oNH «moa cfio moHH NARA «HNH mom muonmfi Hmscme mo zaaasm owma mmm mom whoa non qoc Nae «mm wmvcmEmv Hmuoe wfio cmomfiamgc qmq “Kg nmumfiamfiw wmomfiamfiw mafia umomflamfiw mwmaafiu aflmz Hm mod Hm mm amH mm mm mm ucma cmHkuas Hmsvwmwm NOH «mm mm mm on oaa «q me noumumao xuoaasn cam uouompe mmm mmm moq one qu qu HHN mew :mvmw: msomcmaawumwz chomumav uonma HmscmE How wcmEmo om.mmo oo.o mm.ch oo.noq oo.o oo.o oo.om oo.o Amcv muonma assume he wwumawum cw.- mn.qoa mm.- ww.nm om.on mN.Hm mm.q~ No.Hq «Amnv vcma vmaafiuas Hmsvwmwm co.moN n¢.m©n mm.¢w~ NN.¢~N ww.wcm om.mqm mo.~o~ o~.wom "Away muouomuu >3 wmumamum mm.nmm cc.wm~ nm.mm o~.me ow.mo sq.mm mo.Ho mw.oq "Amsv wxooaasn an umpmamum oo.meH oq.HHHH oo.mow oo.wom oo.mwo oo.omw oo.mmq oo.omq Amsv wcma moan Hmuoa ”vcmH moan mo :OHuuoaoum one max mmH mmeHfi> cw wucmamn Honma Hmscmz .@H xflwcmaa< LIST OF REFERENCES Anonymous, Agricultural Development report, Ministry of Agriculture, Jakarta, 1977. , Agricultural Extension Service, West Java, An annual report, Bandung, 1977. , The second Five Year Development Plan, Ministry of Infor- mation, Jakarta, 1975. , Indonesian Statistics, Bureau of Statistics, Jakarta, 1976, 1977 and 1978. , Feasibility study of tractor development in West Java, Directorate of Production, Pasarminggu, Jakarta, 1977-1978. , The influence of agricultural mechanization upon social and economic progress of the developing countries, The Royal Commonwealth Society Agricultural Conference, London, 1965. , Farm mechanization program for Korea, Exotech Systems, Inc., 1972. , Preliminary evaluation, tractor project South Sulawesi, Indonesia, 1975. , Feasibility study of animal power for agriculture in Java, Report no. 1:19/77/L-77, Agro Economic Survey Institute, Jakarta, 1977. , The consequence of small farm mechanization on rural, employment, incomes and production in selected countries of Asia, A workshOp report, Ag. Engr. Dept., IRRI, September 1978. 125 126 , Meteorology and Geographic Institute, Ministry of Communication, Jakarta, 1975. Arifin, Bustanil, The Indonesian Economy (The struggle for price stabilization), The National Logistics Board, Jakarta, 1976. Bhattacharyya, G. K., Johnson, R. A., Statistical Concepts and Methods, John Wiley and Sons, New York, 1977. Bernstein, Labor shortage as a constraint to increasing cropping intensity, paper presented at the Cropping System Symposium, Los Banos, Philippines, 1980. Binswanger, H. P., The economics of tractors in South Asia, ADC, New York, 1978. Birowo, A. T., Strategy for agricultural mechanization in Indonesia, paper presented at Agricultural Mechanization Seminar, Jakarta, 1977. Bonnen, J. T., Improving information on agriculture and rural life, American Journal of Agricultural Economics 57:753-63, December 1975. Chancellor, W. J. and Singh, 6., Relationship between farm mechaniza- tion and crop yield for farming districts in India, ASAE, Annual Meeting, 1973, paper no. 73-511. Chancellor, W. J., Pattern of tractor management in agricultural development, Univ. of Calif., Davis, 1970. Cochran, W. 6., Sampling Technique, John Wiley and Sons, Inc., New York, 1953. Des Raj, The Design of Sample Survey, McGraw-Hill Book Company, New York, 1972. Duff, B., The economics of small farm mechanization in Asia, paper presented at ORD/ASPAC/FFTC sponsored lecture workshop on the 127 problems of small farm mechanization, Suweon, Korea, 1977. Esmay, M. L., Indonesian selective agricultural mechanization, Directorate of Agricultural Production Seminar, Jakarta, 1978. , A systems approach to rice mechanization in Indonesia, 1971. , Agricultural mechanization and labor utilization in Asia, Agricultural Mechanization in Asia, Summer Edition, 1972. , Systems analysis as a guide for technological transfer, AMA, 1975. , Farming systems approach to appropriate mechanization on small farms, Regional Seminar on Appropriate Mechanization, Jakarta, 1981. , Local manufacturing for small farms in East Asia, Trans- actions of ASAE, 1979. Ferguson, C. F., Micro economic theory, Richard D. Irwin Inc., Homewood, Illinois, 1969. Garmo, D. E. and Canada, J. R., Engineering Economy, Collier-McMillan Publishers, London, 1973. Gaiser, D., Rice post harvest simulation, Ph.D. dissertation, MSU, unpublished, 1980. Gemill, G. and Eicher, C. K., A framework for research on the economics of farm mechanization in developing countries, Michigan State University, 1973. Giles, G. W., The reorientation of agricultural mechanization for developing countries, Shin Norinsha Co., Tokyo, 1975. Gotch, C. B., Technical change and distribution of income in rural areas, economic growth and rural poverty, Southern University, 1972. 128 Hadisapoetro, 8., Expansion of newly developed agricultural land to support the increase of food production, Agronomic Congress, Jakarta, 1977. Khan, A. U., Locally manufactured small farm equipment for Asian farmers, IRRI-Pakistan, Rawalpindi, Pakistan, 1977. Kishu, M., Soil physical properties and machine performances, Tractor and Tillage Machinery Laboratory, Institute of Agricultural Machinery, Omiya, Japan, Jarq. Vol. 6, No. 3, 1972. Kline, C. K. g£_§l,, Agricultural mechanization in equatorial Africa, Institute of International Agriculture, MSU, 1969. Manetsch, T. J. gt al., Systems analysis and simulation with applica- tion to economic and social systems, Part I and II, 1974. Mellor, J. W. g£_§l,, Dilemma on state tube wells, Department of Economics, Cornell University, Ithaca, New York, 1971. , The economics of agricultural development, Cornell Univ. Press, Ithaca, New York, 1974. Mishan, E. J., Cost benefit analysis, new expanded edition, Preager Publishers, New York, 1976. McInerney, J. P. g£_§l,, The consequences of farm tractors in Pakistan, Evaluation of IDA credit for financing the mechanization of farms in Pakistan (1966-1970), IBRD, 1973. McLaughlin, M. M., Agricultural technology transfer and value conflicts, Illinois, 1972. Oldeman, L. R., An agroclimatic of Java, Central Research Institute for Agriculture, Bogor, Indonesia, 1975. 129 Hayami, Y. and Ruttan, W., Agricultural Development and International Perspective, The John Hopkins Press, Baltimore and London, 1971. Steel, R. G. D. and Torrie, J. H., Principles and Procedures of Statistics, McGraw—Hill Book Company, Inc., 1960. Stevens, R. D., Tradition and dynamics in small farm agriculture, Economic Study in Asia, Africa and Latin America, The Iowa State University Press, Ames, Iowa, 1975. Stout, B. A., Selective agricultural mechanization, Seminar at Gadjah Mada University, 1973. Timmer, C. P., Choice in rice milling on Java, ADC, New York, 1974. The World Bank Atlas, 1977. Gadjah Mada University, Soil draft resistance map, 1972-1975. Hunt, D., Farm power and machinery management, Iowa State University Press, Ames, Iowa, 1973. Naylor, T. H. g£_§1,, Computer Simulation Techniques, John Wiley and Sons, New York, 1968. Hillier, F. S. and Lieberman, G. J., Operations research, second edition, Holden-Day, Inc., San Francisco, 1974. wmmmm